The patient survived with a persistent severe neurological disability. A, Lateral projection digital subtraction angiography (DSA) of a left vertebral artery injection demonstrating a high-flow VOGM with posterior choroidal arterial feeders (white arrowhead) draining into an enlarged MPV (black arrowhead). B, Lateral projection DSA of a left vertebral artery injection demonstrating glue cast in the MPV (black arrowhead) with obstruction to outflow, and anterior redirection of outflow of residual shunting through the previously unseen internal cerebral veins (white arrowhead). C, Postprocedure noncontrast axial computed tomography imaging demonstrating hyperattenuating embolic material in the draining venous sac (black arrowhead) and hemorrhagic infarction of the left diencephalon and basal ganglia (white arrowheads). D, Coronal T2 section from magnetic resonance imaging (MRI) performed 3 months following embolization, showing left basal ganglia postinfarction cystic encephalomalacia (black arrowhead) extending into the deep white matter (white arrowheads), suggesting a dependence of venous drainage on the deep system. E, Three-dimensional reconstruction of pre-embolization time-of-flight MRI venography with the benefit of hindsight demonstrates ICV drainage into the venous sac (white arrowheads) despite traditional teaching that the ICVs do not drain into the MPV in patients with VOGM.
eTable. Demographics and treatments for 48 patients with vein of Galen malformations
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Bhatia K, Mendes Pereira V, Krings T, et al. Factors Contributing to Major Neurological Complications From Vein of Galen Malformation Embolization. JAMA Neurol. 2020;77(8):992–999. doi:10.1001/jamaneurol.2020.0825
What are the contributors to major periprocedural neurological complications from endovascular treatment of vein of Galen malformations (VOGMs)?
In this cohort study of 48 children with VOGM, of whom 33 underwent endovascular treatment, 10 of 33 (30%) experienced a major periprocedural neurological complication, half of whom died as a result. The major contributing factors were presence of normal deep venous drainage into the venous sac of the VOGM, excessive embolization of the venous outflow, treatment of more proximal fistulae before treating distal fistulae, and the use of larger microcatheters in neonates.
Pretreatment detailed diagnostic assessment of deep venous outflow, staged transarterial embolization, and initial targeting of distal fistulae should be undertaken to avoid major periprocedural neurological complications in children with VOGM.
Major neurological complications from the embolization of vein of Galen malformations (VOGMs) are poorly understood. We provide a detailed analysis of contributors to periprocedural neurological complications and lessons learned.
To assess the rate of major periprocedural neurological complications following VOGM embolization with major procedural and strategic contributors.
Design, Setting, and Participants
This retrospective cohort study was conducted at a quarternary referral pediatric hospital (Hospital for Sick Children; Toronto, Ontario, Canada) from January 1999 to December 2018 with a mean clinical follow-up of 44.7 months; all children with VOGM diagnosed and/or treated were eligible (n = 48). Thirty-three patients who underwent endovascular treatment were included.
Endovascular staged transarterial embolization performed in 33 patients over 91 sessions.
Main Outcomes and Measures
The primary outcome was the rate of periprocedural neurological complications (occurring within 1 week of embolization). The secondary outcomes were mortality, long-term neurological outcomes, and contributing anatomical and management factors to neurological complications.
Of 33 patients who underwent embolization (31 boys [64.6%]; 17 girls [35.4%]; median age at first embolization, 4 months [range, 0-29 months]), 10 patients (30.3%) developed major periprocedural neurological complications. Five of these patients died. Univariate logistic regression analyses identified internal cerebral vein drainage to the main venous sac of the VOGM and use of a microcatheter with a distal outer diameter of more than 2.0F as significant predictors of poor neurological outcomes. Lessons learned from our experience include the need to assess the internal cerebral vein drainage pattern on preprocedural magnetic resonance venography, avoidance of excessive embolization into the venous sac, treatment of more distal fistulae before proximal fistulae to avoid a sump effect, and preferably use of smaller (<2.0F outer diameter) microcatheters in neonatal embolization procedures.
Conclusions and Relevance
In this cohort, 10 patients with VOGM treated with embolization (30.3%) experienced major periprocedural neurological complications, half of whom died. While these outcomes are superior to historic conservative and surgical treatment results, ongoing improvements in treatment and pretreatment diagnostic approaches are needed. Awareness of the lessons learned from our experience can help to avoid similar complications in the future for this vulnerable population.
Vein of Galen malformations (VOGMs) are congenital intracranial high-flow vascular malformations typically presenting in neonates or infants. They are characterized by arteriovenous shunting from choroidal arteries into a dilated midline deep venous collector: a persistent median prosencephalic vein of Markowski (MPV), the precursor of the future vein of Galen and internal cerebral veins.1,2 These malformations likely result from abnormal morphogenesis of the choroidal vasculature, with formation of arteriovenous fistulous shunts between the eighth and 11th weeks of gestation.1-3 Prenatally, VOGM can be detected in the third trimester on Doppler fetal ultrasonography as a dilated turbulent venous sac posterior to the third ventricle, often associated with high-output cardiomegaly.4 After birth, the 2 common clinical presentations are early neonatal high-output cardiac (and/or multiorgan) failure or later infantile hydrocephalus due to hydrovenous congestion.2,5-7
For neonates with VOGM presenting in cardiac failure, conservative and surgical treatment have mortality rates of greater than 90%.8 Infants without cardiac failure treated surgically have mortality rates of 30% to 40%, with 46% of survivors having significant morbidity.8-10 These unacceptable outcomes led to the development of endovascular treatments that, combined with advances in intensive care and cardiac failure management, have markedly improved the clinical prognosis.2,6,7 Key factors in this improvement have been the purposeful delaying of endovascular treatment until the third to sixth month of life when possible, the use of a staged approach aiming to progressively reduce arteriovenous shunting, and centralization of care in dedicated pediatric centers with multidepartmental and pediatric neurovascular expertise.2,6
Lasjaunias et al,6 who published the largest series to our knowledge to date (n = 216) of children treated by an endovascular approach, used the previously mentioned principles and reported a mortality of 52% in neonates (12/23), 7.2% in infants (11/153), and 0% in older children (0/40). A meta-analysis of clinical outcomes after endovascular treatment reported an overall mortality of 14% (27% in neonates and 1% in infants; P < .001), poor neurological outcomes in 21% (22%/16%), good neurological outcomes in 62% (48%/77%; P < .001), and complete occlusion of the VOGM in 56% (59%/56%).7 Similar historical outcomes were also previously published by our group11 and by Berenstein et al.12 These results demonstrate that outcomes with endovascular treatment are superior compared with historic data from surgical approaches but are still associated with mortality or morbidity in approximately one-third of cases, with the worst outcomes among patients requiring emergent treatment in the neonatal period.2,6,7
While overall complication and periprocedural hemorrhage rates with VOGM embolization have been reported via descriptions of numeric incidence,6,7,11,12 to our knowledge few studies to date have analyzed anatomical and treatment strategy factors that contribute to major neurological complications.5 Given the few centers worldwide regularly performing VOGM embolization and the potential for selection or publication bias in the current literature (in which case reports/series predominate), lessons learned from poor outcomes may not be shared among the wider medical community.
This study aims to assess anatomical and management factors that contributed to major periprocedural neurological complications following VOGM embolization. The analysis of these contributory anatomical and management factors extends beyond the status of the predominantly descriptive existing literature. We hope that the lessons learned will aid in directing individualized treatment strategies and ultimately in improving outcomes.
Ethics approval, including a waiver for obtaining informed consent from participants given the retrospective nature of the study and the long time period being assessed, was granted by the research ethics board of the Hospital for Sick Children, Toronto, Ontario, Canada. This study was a retrospective cohort study of all patients with VOGM managed at the Hospital for Sick Children over a 20-year period from January 1999 to December 2018. The primary outcome was the rate of periprocedural neurological deterioration, defined as a sustained reduction in global or focal neurological function compared with the preprocedural baseline occurring within 1 week of an embolization procedure. The secondary outcomes were overall mortality, periprocedural mortality (within 1 week of embolization), access site–associated complications, immediate embolization angiographic results, long-term embolization angiographic results, and long-term neurological outcomes (at the last follow-up using the electronic patient records). Subgroup analyses were performed by emergent vs elective embolization and age at first embolization (<4 months, 4-12 months, and >12 months).
The primary and secondary outcomes were assessed using data obtained from the clinical records and imaging (via picture archiving and communication system) by 2 independent reviewers (K.B. and P.K.) who were fellowship-trained interventional neuroradiologists. Disagreements were resolved by consensus; otherwise they were resolved by the senior interventional neuroradiologist overseeing this study (T.K.). This same process applied to determining whether a poor clinical outcome was associated with the embolization procedure (ie, a complication).
All available baseline and follow-up magnetic resonance imaging (MRI), computed tomography (CT), and digital subtraction angiography (DSA) imaging studies for each patient were assessed independently by the 2 reviewers. Imaging characteristics sought from MRI, CT, and DSA results included angiographic classification (eg, choroidal, mural, or mixed type), dural venous thrombosis, dural venous congestion, tonsillar prolapse, global or focal infarct/encephalomalacia, intracranial hemorrhage, internal jugular vein stenosis, and the number of embolization sessions, specific arteries supplying the VOGM at baseline, and nature of the venous drainage at baseline.
The embolic agent(s) used, endovascular approach (transarterial, transvenous, or both), vascular access site, sheath size, microcatheter(s) used, number of arterial feeders embolized, immediate embolization angiographic result (eg, cure, partial reduction, or no change), and long-term embolization results were identified. For patients who experienced a periprocedural neurological deterioration, 4 additional factors were analyzed: (1) treatment strategy, (2) intraprocedural technical challenges, (3) drainage pathways for the internal cerebral veins (ICVs), and (4) age at the time of the complication.
Descriptive analyses were performed on the agreed outcome results. Univariate logistic regression analyses were performed to assess for predictors of poor long-term neurological outcomes (defined as moderate-disability hemiparesis requiring assistance with mobilization, severe disability–severe cognitive or motor deficits, or death) using the binary predictor variables of male sex, younger than 4 months at the time of the first embolization procedure, immediate angiographic cure at the end of an embolization procedure, use of microcatheter with a distal outer diameter (OD) of more than 2.0F, and the presence of unilateral or bilateral ICV drainage into the venous sac of the VOGM (σ = .05; StatsDirect, version 3.1.22 [Tidestone Technologies Inc]). Univariate analyses were favored because of the likely low number of outcome events.
Of 48 patients with VOGM (31 boys [64.6%]; 17 girls [35.4%]) treated at the Hospital for Sick Children, between 1999 and 2018, 33 (68.8%) underwent endovascular transarterial glue embolization over a total of 91 sessions (μ = 2.8 sessions per patient; σ = 2.04) (eTable in the Supplement). Of the 15 patients who did not undergo embolization, 9 (60.0%) received palliative care (5 because of global parenchymal brain damage on baseline MRI and 4 because of multiorgan failure with a Bicetre neonatal score <8), 3 (20.0%) had spontaneous thrombosis of the malformation, 2 (13.3%) were asymptomatic at presentation, and 1 (6.7%) underwent surgical clipping at an external center. Age at diagnosis, treatment(s) undertaken, age at first embolization, mortality rates, and angiographic classification of the VOGMs are summarized in eTable in the Supplement. For survivors, mean imaging follow-up for was 39.2 months (σ = 13.1 months) and clinical follow-up was 44.7 months (σ = 17.1 months).
Of the 33 patients who underwent 91 embolization sessions, 10 patients (30.3% of patients, 11% of embolization sessions) experienced a periprocedural neurological complication, of whom 5 died within 1 week of the procedure (15% of patients who underwent embolization, 50% of patients with a periprocedural neurological complication) (Table 1). All 5 of these deaths resulted from the neurological complication. Of the 10 neurological complications, 6 patients experienced a deep venous infarct with hemorrhagic transformation, 2 patients had nontarget arterial embolization of glue resulting in arterial infarcts, and 2 neonates had intraprocedural arterial perforations (both fatal). Four of 13 patients (31%) younger than 4 months (all done on an emergent basis for progressive cardiac failure), 5 of 16 patients (31%) aged 4 to 12 months, and 1 of 4 patients (25%) older than 12 months at first embolization had a major neurological complication. Major nonneurological complications (n = 3) are also included in Table 1.
Death occurred in 6 of 33 patients who underwent embolization (18%). Of these 6 deaths, 1 death (16.7%) occurred because of rapid deterioration of cardiac failure while awaiting repeated embolization (this patient was initially treated on an emergent basis at age 3 months) and 5 deaths (83.3%) occurred because of periprocedural major neurological complications (5 of 48 patients [10%]; 5 of 33 patients who underwent embolization [15%]). Two (40%) of these deaths were in patients undergoing their first embolization on an emergent basis at younger than 4 months (both arterial perforations), 2 (40%) were in patients who first underwent embolization at age 4 to 12 months on a semielective basis (both deep venous infarcts), and 1 (20%) occurred in a patient who first underwent embolization at age 29 months on a semielective basis (deep venous infarct). The remaining secondary outcomes for all patients who underwent embolization are detailed in Table 2.
Univariate logistic regression analyses for predictors of poor long-term neurological outcomes are detailed in Table 3. Two variables were found to be statistically significant predictors of poor outcome in this series: (1) use of a microcatheter with a distal OD of more than 2.0F (odds ratio, 7.11; 95% CI, 1.40-36.12; P = .02) and (2) presence of unilateral or bilateral deep venous/ICV drainage to the primary venous sac of the VOGM (odds ratio, 24.00; 95% CI, 2.47-233.45; P = .01).
The contributing factors and lessons learned from the major neurological complications are outlined in Table 4. The major identified contributing factors to deep hemorrhagic venous infarction were: (1) ICV drainage directly into the MPV, (2) occlusion of the MPV/venous sac outflow before complete embolization, and (3) rapid thrombosis of the venous drainage outlet with angiographic cure of the VOGM after a single session of treatment. For nontarget arterial embolization, the major identified contributing factor was a sump shunting effect from distal fistulae when treating a more proximal fistula first. For intraprocedural arterial perforations, both patients were neonates undergoing emergent embolization of posterior choroidal arterial feeders via the basilar artery. In both cases, the microcatheter being used had a distal OD of more than 2.0 F (>0.67 mm).
Endovascular transarterial embolization is the standard of care for patients with symptomatic VOGM.2,6,13 However, despite superior outcomes compared with conservative or surgical treatment, approximately one-third of treated patients will die or have significant long-term neurological disability following embolization.6,7,11,12 There are numerous small case series in the literature outlining successful treatment7 and several larger cohorts (>10 patients, largest sample size of 216 patients) also published,5,6,11,12,14 with the latter providing overall mortality and complication rates over approximately 10 to 20 years at single centers.
Lasjaunias et al6 described their use of the Bicetre neonatal score (scored out of 21 points assessing neonatal baseline function of cardiac, neurological, respiratory, renal, and hepatic systems) to prognosticate outcomes and aid in decision-making.6 Under their management approach, a score of less than 8 (multi-organ failure) would be a relative contraindication to treatment because of a likely poor long-term outcome, 8 to 12 would warrant emergency endovascular treatment, and more than 12 would allow for medical management and deferral of endovascular treatment until the lower risk treatment age window of 4 to 6 months.6 A meta-analysis of outcomes following VOGM embolization found that studies using the Bicetre neonatal score for patient selection were associated with superior rates of good neurological outcomes compared with those that did not (of course this is affected by selection bias).7
Geibprasert et al5 identified clinical and imaging features associated with good neurological outcomes following embolization, including asymptomatic clinical presentation, a Bicetre neonatal score of more than 12, absence of brain parenchymal damage on baseline MRI results, absence of arterial steal on angiography results, and lower numbers of arterial feeders.5 Logistic regression analyses in our series identified use of a more than 2.0 F distal OD microcatheter and presence of ICV drainage to the MPV as significant predictors of poor neurological outcomes.
Using meta-analysis data, technical complications from endovascular treatment of VOGM occur in 19% of patients (29% in neonates/10% in infants), perioperative hemorrhage in 9% (12%/4%), and perioperative ischemia in 1% (3%/0%).7 Meta-analysis results have also described overall postembolization mortality rates of 10% and complication rates of 37%.15 Our periprocedural neurological complication rate (10 of 33 patients [30.3%]) is contributed to by a combination of deep cerebral hemorrhagic infarction (6 [60.0%]), nontarget arterial glue embolization (2 [20.0%]), and intraprocedural arterial perforation (2 [20.0%]). Our overall combined neurological and nonneurological postembolization mortality (18%) and complication (39%) rates are comparable with published meta-analyses on the subject (10% and 37%, respectively),15 and therefore are likely representative of the true treatment risk. Later in this article, we highlight lessons learned from these complications (Table 4).
There is increasing awareness of a subset of patients who experience hemorrhagic infarction of the diencephalon and striatum in the hours that follow embolization for VOGM.2,12 The mechanism of this complication is thought to be venous infarction because of the occlusion of outflow from the ICVs (or equivalent deep venous drainage system that may, contrary to previous beliefs, be dependent on the MPV).2,6,12,16 Secondary hemorrhagic transformation of the venous infarction is likely due to normal perfusion pressure breakthrough in ischemic tissue resulting from rapid obliteration of the high-flow shunt.2,16 Lasjaunias et al6 also reported 2 cases of postprocedural hemorrhage in the deep cerebral structures after transvenous occlusion of the draining sac.6 This mechanism may explain poor outcomes seen in cases using a transvenous approach in which the entire draining venous sac was occluded13 or in cases for which transarterial embolization excessively extended into the venous sac.2,6
In our cohort, 6 patients (18.2%) experienced periprocedural hemorrhagic infarction of the deep cerebral structures (Table 2). In one example case, communication of the ICVs with the MPV was demonstrated on DSA results during the embolization process (Figure), leading to an emergent halting of the procedure. Unfortunately, the patient still experienced a hemorrhagic infarction of the diencephalon despite a single-session angiographic cure of the VOGM, surviving with a persistent severe neurological deficit. With the benefit of hindsight, the retrospective review of the preprocedural MRI venogram was suggestive of this anatomical pattern (Figure). In another patient, there was inadvertent glue migration into the torcula and transverse sinus, resulting in venous outflow obstruction and death, again following a single-session cure of the VOGM. In the remaining 4 patients, excess venous extension of glue and/or rapid thrombosis of the venous sac may have been contributing factors (with angiographic cure of the VOGM after a single session of treatment in 2 cases). Four of these 6 patients (66.7%) had an angiographic cure of their VOGM after a single treatment session, and in all cases, obstruction of venous outflow to the deep cerebral structures was the likely cause of the subsequent hemorrhagic infarction.
Traditional teaching regarding the embryology of VOGMs has suggested that the vein of Galen normally develops from the posterior third of the MPV (with the anterior two-thirds involuting and forming the connection to the ICVs) and the deep venous drainage of the diencephalon and striatum do not gain access into the persistent enlarged venous sac of a VOGM. Rather, it was thought the ICVs drain via accessory pathways, such as the lateral mesencephalic vein toward the superior petrosal sinus (resulting in the epsilon sign on angiography).6 However, Raybaud et al3 (the first to postulate that the venous collector in VOGM was the MPV and not the true vein of Galen) identified drainage of the ICVs into the MPV in 4 of 12 patients.3 Drainage of the ICVs to the MPV in some patients with VOGM has also been suggested by multiple authors since then using magnetic resonance venography, delayed-phase cerebral angiography, and retrograde transvenous angiography.2,17,18 More recently, time-of-flight magnetic resonance venography assessment (associated with cerebral angiography) of 55 patients with VOGM identified direct drainage of the ICVs to the venous sac in 15 patients (27.3%).16 A logistic regression analysis in our series identified deep venous drainage to the MPV as a statistically significant predictor of poor outcomes after embolization (Table 3).
There is convincing evidence from multiple studies demonstrating that in a larger than previously thought subset of patients with VOGM the deep veins can drain into the MPV and MRI time-of-flight venography can be useful to help identify such drainage.1-3,5,16-18 Identifying these patients may prevent catastrophic outcomes. We presume that the reason why this phenomenon has not been more widely recorded on angiography is associated with several factors: competitive flow from the fistulous components of the VOGM, arterial-phase predominant angiographic acquisitions after injection of contrast to minimize radiation dosage, contrast recirculation in the arteries masking venous channels, predominant focus of interventionists on the fistulae, and technological limitations of older angiographic equipment.1,2
In the light of these findings, we make 3 recommendations. First, routine detailed high-resolution MRI venographic assessment before embolization should be undertaken to assess the drainage pattern of the ICVs; second, 3-vessel angiography (internal carotid arteries and the dominant vertebral artery, held into the late venous phase) be performed in all cases before commencing embolization to identify the ICVs and help avoid excessive penetration of embolic material into the venous sac; and third, staged embolization strategies should be used to minimize the chance of rapid thrombosis of the venous sac (Table 4).
Our results, and those of prior authors,2,3,6,12 suggest that VOGMs should not be approached in the same manner as adult intracranial fistulous lesions, for which occlusion of the venous drainage is typically curative and considered part of the treatment goal. Rather, rapid occlusion of the venous sac in patients with VOGM can result in disastrous neurological complications, despite angiographic cure of the malformation. For this reason, transvenous embolization should be avoided unless there is no viable transarterial route remaining.2,6 Staged progressive transarterial reduction of the shunts (as proposed by Lasjaunias et al6) with the goal of achieving a normally developing child is therefore preferable to a rapid angiographic cure of the lesion.
Two patients in the cohort experienced nontarget arterial glue embolization (Table 2). In both cases, a sump shunting effect was seen when transarterial glue embolization was being undertaken via posterior choroidal feeders arising from the posterior cerebral arteries and there was coexistent shunting from more distal pericallosal fistulous feeders (arising from the anterior cerebral arteries) to the VOGM. This resulted in nontarget glue embolization across the posterior communicating arteries and into the anterior circulation via a sump effect. Both patients survived despite large cerebral hemisphere infarcts but were left with persistent hemiparesis. Based on these 2 cases (and similar experiences we have had in children with multiple intracranial high flow shunts in the setting of hereditary hemorrhagic telangiectasia), we recommend initial treatment of the more distal pericallosal fistulous feeders or major mural fistulous feeders to the VOGM before embolization of more proximal posterior choroidal feeders.
Two patients experienced an intraprocedural arterial perforation (fatal in both cases). Both patients were neonates being treated in their first week of postnatal life on an emergent basis for progressive cardiac failure. In both cases, embolization was being undertaken with access across the basilar artery to embolize the VOGM via posterior choroidal artery feeders. In this very young age group being treated in an emergent setting, the small caliber and fragility of the vasculature may have been contributing factors.
An additional potential contributor was the relatively large size of microcatheters used in both cases (2.3 F and 2.4 F distal OD, respectively). The logistic regression analysis in our series identified the use of a microcatheter with a distal OD of more than 2.0 F as a statistically significant predictor of poor outcomes. Given that there are multiple smaller diameter microcatheters available for neurovascular procedures (eg, 1.2–1.9 F distal OD) that can be navigated in a flow-directed fashion over a softer 0.007/0.008-in guidewire, the use of larger-diameter microcatheters in neonatal embolization procedures should be reconsidered. An extensive discussion of nonneurological complications from VOGM embolization in our cohort is beyond the scope of this article (the complications are summarized in Table 1).
This study is limited by its retrospective nature and its confinement to a single center. Shared techniques between the 3 neurointerventionists involved in these procedures may have affected outcomes. To date in the literature, prospective or multicenter data on the management of VOGM is lacking and may be best addressed through a multicenter prospective registry.
Major periprocedural neurological complications occurred in approximately one-third of patients who underwent endovascular treatment for VOGM in our cohort, half of whom died as a consequence of the complication. Lessons learned from our experiences suggest that deep venous drainage patterns need to be diligently assessed, excessive single-session embolization into the venous sac should be avoided, distal fistulous feeders should be embolized before more proximal feeders, and smaller diameter microcatheters (<2.0 F) may be preferable for neonatal procedures.
Accepted for Publication: February 21, 2020.
Corresponding Author: Kartik Bhatia, MBBS, MS, PhD, Department of Medical Imaging, Sydney Children’s Hospital Network, Cnr Hawkesbury Rd and Hainsworth St, Westmead, NSW 2145, Australia (firstname.lastname@example.org).
Published Online: April 27, 2020. doi:10.1001/jamaneurol.2020.0825
Author Contributions: Drs Bhatia and Muthusami (principal investigator) had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Bhatia, Krings, Dirks, Armstrong, Muthusami.
Acquisition, analysis, or interpretation of data: Bhatia, Mendes Pereira, Krings, Terbrugge, Kortman, Shroff, Muthusami.
Drafting of the manuscript: Bhatia, Mendes Pereira, Kortman, Armstrong, Muthusami.
Critical revision of the manuscript for important intellectual content: Bhatia, Mendes Pereira, Krings, Terbrugge, Dirks, Shroff, Muthusami.
Statistical analysis: Bhatia.
Administrative, technical, or material support: Kortman, Shroff, Muthusami.
Supervision: Mendes Pereira, Krings, Terbrugge, Dirks, Shroff, Muthusami.
Conflict of Interest Disclosures: Dr Krings reported personal fees from Medtronic, Stryker, Penumbra, Thieme, and Marblehead outside the submitted work. Dr Shroff reported offering occasional medical expert testimony, serving as a speaker for the medical advisory board of BioMarin, and receiving speaking fees for the Speaker for Innovations Update in Pediatric Neuroradiology (November 2019; Sao Paolo, Brazil) for Siemens. No other disclosures were reported.
Create a personal account or sign in to: