Abstract
Introduction: There has been rapid advancement in the development of deep brain stimulation (DBS) as a treatment option for adults for neurological and neuropsychiatric conditions. Here, we present a scoping review of completed and ongoing clinical trials focused on DBS in pediatric populations, highlighting key knowledge gaps. Methods: Three databases (PubMed, OVID, and Embase) and the clinicaltrials.gov registry were queried to identify clinical trials for DBS in pediatric cohorts (age ≤18 years). Prospective and retrospective case series were excluded. No restrictions were placed on the diagnoses or measured clinical outcomes. Individual patient demographics, diagnosis, DBS target, and primary endpoints were extracted and summarized. Results: A total of 13 clinical trials were included in the final review, consisting of 9 completed trials (357 screened) and 4 ongoing trials (82 screened). Of the completed trials, 6 studied dystonia (both inherited and acquired; participants aged 4–18 years) and 3 studied drug-resistant epilepsy (participants aged 4–17 years). Among the 6 trials for dystonia, 5 used the Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS) as the primary endpoint. There were a total of 18 adverse events documented across 63 participants, with 5 of 9 studies reporting adverse events. Ongoing clinical trials are evaluating DBS for dystonia (N = 2), epilepsy (N = 1), and self-injurious behavior (N = 1). Conclusions: This scoping review summarizes the landscape of clinical trials for DBS in children and youth. In dystonia, further research is warranted with more relevant pediatric outcome measures and for understudied patient subgroups and targets. There are also significant gaps in our understanding of evaluating the role of DBS in other neurological and neurodevelopmental disorders in pediatric populations.
Introduction
Deep brain stimulation (DBS) has become a mainstay treatment for adult movement disorders such as Parkinson’s disease and essential tremor [1, 2]. In adults, DBS has also been subject to at least phase I clinical trials in many other disorders, including neuropsychiatric conditions (e.g., anorexia nervosa [3], obsessive-compulsive disorder [4], and major depressive disorder [5]), traumatic brain injury [6], disorders of consciousness [7], and obesity [8], among others. A recent review identified nearly 400 clinical trials spanning 28 different disorders and 26 brain targets in adult populations [9]. Nearly 40% of these were indications other than movement disorders, including psychiatric disorders, epilepsy, and addiction-related disorders.
DBS is considered off-label and/or investigational for all pediatric indications, with the exception of dystonia for which there is an FDA Humanitarian Device Exemption for children who are 7 years and older. The applications of DBS for disorders that present primarily in childhood beyond dystonia [10‒12] remain controversial. For example, although DBS has been successfully used for Tourette’s syndrome in children [13], data are based on open-label studies, which are prone to bias. In the example of Tourette’s syndrome, several groups have advocated for its disuse in pediatric cohorts due to the possibility of self-remission of symptoms with age [14‒16]. Furthermore, children represent an intrinsically vulnerable population due to concerns regarding informed consent, autonomy, capacity, and long-term consequences of invasive treatments [17]. This may temper enthusiasm for conducting clinical trials in pediatric populations. Therefore, the application of DBS in children and youth poses numerous medical, bioethical, and regulatory challenges.
Much of the published literature comprises limited case series with few dedicated trials in pediatric cohorts [18]. It is challenging for adult DBS literature to translate directly to children for many reasons. First, the primary outcome measures used in adult trials may be less valid in pediatrics. For example, the Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS), a commonly used primary outcome measure, is influenced by age and longitudinal assessments for the same child may be confounded by physiologically immature movements and posturing [19]. Furthermore, measures in adults may not be the most relevant in a pediatric population. While BFMDS has not showed a significant change following DBS for acquired dystonia in children, other meaningful pediatric-specific outcome measures may be influenced, including pain and comfort, school attendance, seating tolerance, and caregiver burden [20]. Other considerations in children also warrant specific studies in this population, including ongoing brain development and timing of surgery (i.e., anatomical size differences), differential risks of DBS procedures in children (e.g., anesthetic considerations) [21], and neurodevelopmental diagnoses in childhood that may be amenable to neuromodulation during critical windows for brain development (e.g., cerebral palsy) [22]. The lack of robust evidence of applying DBS to pediatric conditions is often cited as a prohibitive factor in considering DBS therapy in this vulnerable population [17].
This scoping review aims to characterize the clinical trial landscape of pediatric DBS by reviewing completed and ongoing trials in children and youth up to age 18 years. We found a disproportionate representation of DBS studies for dystonia and highlight the gaps in knowledge due to limited clinical research for other pediatric neurological and neurodevelopmental disorders.
Methods
Search Strategy
A literature search was conducted of articles available across three electronic databases: PubMed, OVID, and Embase, published before July 7, 2024. This search was completed per PRISMA-ScR guidelines (Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews) [23]. Search strategies were refined in collaboration with an institutional librarian, with the following terms for PubMed: ([pediatric] OR [children] OR [youth]) AND ([“deep brain stimulation”] OR [“responsive neurostimulation”] OR [“neuromodulation”] OR [“DBS”]). These terms were adjusted to meet the requirements of the other databases. The search included inception to July 7, 2024. No publication status, year, or language restrictions were made to capture the full scope of the literature. A PRISMA schematic provides an overview of the search process (Fig. 1). A separate search for active clinical trials of pediatric DBS, which captured studies not yet published in PubMed/OVID/Embase, was performed using the publicly available trial registry, ClinicalTrials.gov (https://www.clinicaltrials.gov/), maintained by the US National Library of Medicine. Where possible, reported endpoints were synthesized in a meta-analysis to determine outcomes after DBS. The “metafor” function from the “meta” package of RStudio version 2023.09.1 (RStudio Inc., Boston, MA, USA) was used to conduct a pooled meta-analysis of proportions for studies reporting outcomes in respect to the BFMDS and evaluate the heterogeneity across the included studies.
PRISMA flow diagram indicating the number of published records identified, screened, eligible, and included in the review.
PRISMA flow diagram indicating the number of published records identified, screened, eligible, and included in the review.
Screening and Eligibility Criteria
After removing duplicate studies, two authors (K.M. and Y.J.) performed an independent initial screen of the abstracts to exclude irrelevant articles. This was followed by a full-text review of all remaining articles (K.M. and Y.J.). Discrepancies were reviewed and settled by a third author (H.S.). We included all clinical trials and prospective studies evaluating DBS in pediatric patients for any indication of whether the primary outcome was a clinical response to DBS. We excluded retrospective studies, case series, case reports, and studies primarily recruiting adult patients or using another form of neuromodulation (e.g., transcranial magnetic stimulation). A comprehensive screening checklist is provided in online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000543289).
Data Extraction and Analysis
Collected data included patient demographics (age, sex), clinical diagnosis, disease duration, DBS target, primary endpoint (clinical rating scale), DBS response (pre- vs. post-op clinical scores), DBS duration until follow-up, and reported adverse events (Table 1). Two authors (Y.J., K.M.) independently determined a risk of bias assessment using the Risk Of Bias In Nonrandomized Studies – of Interventions (ROBINS-I) tool [24], which consisted of 7 domains of potential bias (i.e., confounding, participant selection, classification of interventions, deviation from intended interventions, missing data/attrition, outcome measurements, and reporting) and resulted in an overall bias score.
Published clinical trials of DBS for pediatric populations
Diagnosis . | Study (first author, publication year, location) . | Study design . | N . | Age (years) . | Sex (% male) . | Disease duration (years) . | DBS target . | Primary endpoint (clinical rating scale) . | Surgical techniques . | DBS response (pre- and post-DBS clinical scores, %improvement with DBS) . | DBS duration (follow-up in months) . | Adverse events . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Cerebral palsy (dyskinetic) | Koy et al. [25], 2022, Germany | RCT | 16 | 13.4±2.9 | 62.5 | n/a | GPi (bilateral) | Caregiver Priorities & Child Health Index of Life with Disabilities (CPCHILD) | DBS implantation as per institutional protocols | Pre-DBS: 46.5±14.4 | 12 | Infection (n = 1), ICH (n = 1), programming/device use error (n = 2), worsening symptoms (n = 1), fatigue (n = 1), headache (n = 2), hypersalivation (n = 1), seroma (n = 1), scar pain (n = 1) |
Frequency: 90–130 Hz | Post-DBS: 50.7±17.0 | |||||||||||
Pulse width: 90–150 μs | Improvement: 8.7±21.8% | |||||||||||
Amplitude: 0.2 mA | ||||||||||||
Dystonia (generalized) | Borggraefe et al. [26], 2010, Germany | Prospective | 6 | 9±2.2 | 50 | 5.3±4.1 | GPi (bilateral) | BFMDRS | MRI-guided stereotaxy using Leksell/Lerch system | Pre-DBS: 51.2±14.0 | 14.8±3.2 | None reported |
Permanent implantation after 5 days of testing | Post-DBS: 14.0±11.7 | |||||||||||
Improvement: 74.9±16.7% | ||||||||||||
Dystonia | Candela et al. [27], 2018, Spain | Prospective | 6 | 12±3.3 | 50 | 7.2±3.9 | GPi (bilateral) | BFMDRS | Preoperative MRI used to identify coordinates for GPi target. Head stabilized using a robotic arm head-holder. Starting parameters 60 μs, 130 Hz, and adjusted accordingly | Pre-DBS: 44±22 | 6 | Electrode migration (n = 1), transient dysarthria (n = 1), transient gait instability (n = 1) |
Post-DBS: 17±16 | ||||||||||||
Improvement: 61.8±26.4% | ||||||||||||
Dystonia | Legros et al. [28], 2004, France | Prospective | 10 | 12±3.4 | 40 | n/a | GPi (bilateral) | BFMDRS | Electrodes placed based on MRI-determined coordinates. Parameters set to 450 μs and 130 Hz, amplitude 0.8A and 2.4V | Pre-DBS: 58.6±29.2 | 9.0±1.1 | Not specified |
Post-DBS: 44±23.7 | ||||||||||||
Improvement: 24.4±16.5 | ||||||||||||
Dystonia (secondary, from perinatal brain injury (N = 3), encephalitis (N = 1)) | San Luciano et al. [29], 2020, USA | Phase I clinic trial | 4 | 13.0±2.9 | 50 | 11.0±3.6 | Vop/Vim (bilateral) | BFMDRS | DBS implantation as per institutional protocol, using a skull-mounted aiming device. Parameters set to 60 ms, 130 Hz, 0–5 V | Pre-DBS: 54.5±10.3 | 12 | None reported |
Post-DBS: 58.7±19.6 | ||||||||||||
Improvement: 1.6±13.6% | ||||||||||||
Dystonia (primary) | Starr et al. [30], 2014, USA | Prospective | 6 | 11.0±2.8 | 67 | 3.0±1.8 | GPi (bilateral) | BFMDRS | Skull-mounted framing device with proprietary ClearPoint software used. Head positioned in 4-point head holder, with reduced pressure applied for children <10 years | Pre-DBS: 39.5±16.1 | 12 | Transient neurological deficit |
Post-DBS: 4.9±5.7 | ||||||||||||
Improvement: 75.4±12.5 | ||||||||||||
Epilepsy | Chabardès et al. [31], 2002, France | Prospective | 3 | 10.3±5.0 | 67 | 7.5±3.3 | STN (bilateral) | Seizure frequency | DBS implantation as per institutional protocols; radionics stereotactic frame used for positioning, electrodes implanted based on ventriculography-based coordinates | 2/3 patients had a significant reduction in seizure frequency | 17.0±9.9 | Infection (n = 1) |
Epilepsy | Suresh et al. [32], 2024, Canada | Randomized control trial, patient preference | 18 | 13.9±2.7 (VNS), 14.8±2.7 (add on DBS) | 55.6 | n/a | CMN or anterior nucleus (bilateral) | Seizure frequency | DBS implantation as per institutional protocol. Parameters set to 90 μs and 145 Hz, amplitude 2.2–5 mA | Significant reduction in seizure frequency (51.9% vs. 12.3%, p = 0.047) | 12 | None reported |
Epilepsy (Lennox-Gastaut syndrome) | Velasco et al. [33], 2006, Mexico | Prospective | 9 | 9.8±2.9 | n/a | 7.4±6.7 | CMN thalamus (bilateral) | Seizure frequency | DBS implantation as per institutional protocol. Parameters set to 0.45 ms, 130 Hz | 6/9 participants experienced reduced seizure frequency | 18 | Skin erosion requiring explanation (n = 2) |
Diagnosis . | Study (first author, publication year, location) . | Study design . | N . | Age (years) . | Sex (% male) . | Disease duration (years) . | DBS target . | Primary endpoint (clinical rating scale) . | Surgical techniques . | DBS response (pre- and post-DBS clinical scores, %improvement with DBS) . | DBS duration (follow-up in months) . | Adverse events . |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Cerebral palsy (dyskinetic) | Koy et al. [25], 2022, Germany | RCT | 16 | 13.4±2.9 | 62.5 | n/a | GPi (bilateral) | Caregiver Priorities & Child Health Index of Life with Disabilities (CPCHILD) | DBS implantation as per institutional protocols | Pre-DBS: 46.5±14.4 | 12 | Infection (n = 1), ICH (n = 1), programming/device use error (n = 2), worsening symptoms (n = 1), fatigue (n = 1), headache (n = 2), hypersalivation (n = 1), seroma (n = 1), scar pain (n = 1) |
Frequency: 90–130 Hz | Post-DBS: 50.7±17.0 | |||||||||||
Pulse width: 90–150 μs | Improvement: 8.7±21.8% | |||||||||||
Amplitude: 0.2 mA | ||||||||||||
Dystonia (generalized) | Borggraefe et al. [26], 2010, Germany | Prospective | 6 | 9±2.2 | 50 | 5.3±4.1 | GPi (bilateral) | BFMDRS | MRI-guided stereotaxy using Leksell/Lerch system | Pre-DBS: 51.2±14.0 | 14.8±3.2 | None reported |
Permanent implantation after 5 days of testing | Post-DBS: 14.0±11.7 | |||||||||||
Improvement: 74.9±16.7% | ||||||||||||
Dystonia | Candela et al. [27], 2018, Spain | Prospective | 6 | 12±3.3 | 50 | 7.2±3.9 | GPi (bilateral) | BFMDRS | Preoperative MRI used to identify coordinates for GPi target. Head stabilized using a robotic arm head-holder. Starting parameters 60 μs, 130 Hz, and adjusted accordingly | Pre-DBS: 44±22 | 6 | Electrode migration (n = 1), transient dysarthria (n = 1), transient gait instability (n = 1) |
Post-DBS: 17±16 | ||||||||||||
Improvement: 61.8±26.4% | ||||||||||||
Dystonia | Legros et al. [28], 2004, France | Prospective | 10 | 12±3.4 | 40 | n/a | GPi (bilateral) | BFMDRS | Electrodes placed based on MRI-determined coordinates. Parameters set to 450 μs and 130 Hz, amplitude 0.8A and 2.4V | Pre-DBS: 58.6±29.2 | 9.0±1.1 | Not specified |
Post-DBS: 44±23.7 | ||||||||||||
Improvement: 24.4±16.5 | ||||||||||||
Dystonia (secondary, from perinatal brain injury (N = 3), encephalitis (N = 1)) | San Luciano et al. [29], 2020, USA | Phase I clinic trial | 4 | 13.0±2.9 | 50 | 11.0±3.6 | Vop/Vim (bilateral) | BFMDRS | DBS implantation as per institutional protocol, using a skull-mounted aiming device. Parameters set to 60 ms, 130 Hz, 0–5 V | Pre-DBS: 54.5±10.3 | 12 | None reported |
Post-DBS: 58.7±19.6 | ||||||||||||
Improvement: 1.6±13.6% | ||||||||||||
Dystonia (primary) | Starr et al. [30], 2014, USA | Prospective | 6 | 11.0±2.8 | 67 | 3.0±1.8 | GPi (bilateral) | BFMDRS | Skull-mounted framing device with proprietary ClearPoint software used. Head positioned in 4-point head holder, with reduced pressure applied for children <10 years | Pre-DBS: 39.5±16.1 | 12 | Transient neurological deficit |
Post-DBS: 4.9±5.7 | ||||||||||||
Improvement: 75.4±12.5 | ||||||||||||
Epilepsy | Chabardès et al. [31], 2002, France | Prospective | 3 | 10.3±5.0 | 67 | 7.5±3.3 | STN (bilateral) | Seizure frequency | DBS implantation as per institutional protocols; radionics stereotactic frame used for positioning, electrodes implanted based on ventriculography-based coordinates | 2/3 patients had a significant reduction in seizure frequency | 17.0±9.9 | Infection (n = 1) |
Epilepsy | Suresh et al. [32], 2024, Canada | Randomized control trial, patient preference | 18 | 13.9±2.7 (VNS), 14.8±2.7 (add on DBS) | 55.6 | n/a | CMN or anterior nucleus (bilateral) | Seizure frequency | DBS implantation as per institutional protocol. Parameters set to 90 μs and 145 Hz, amplitude 2.2–5 mA | Significant reduction in seizure frequency (51.9% vs. 12.3%, p = 0.047) | 12 | None reported |
Epilepsy (Lennox-Gastaut syndrome) | Velasco et al. [33], 2006, Mexico | Prospective | 9 | 9.8±2.9 | n/a | 7.4±6.7 | CMN thalamus (bilateral) | Seizure frequency | DBS implantation as per institutional protocol. Parameters set to 0.45 ms, 130 Hz | 6/9 participants experienced reduced seizure frequency | 18 | Skin erosion requiring explanation (n = 2) |
BFMDRS, Burke-Fahn-Marsden Dystonia Rating Scale; CMN, centromedian nucleus; GPi, globus pallidus internus; ICH, internal cerebral hemorrhage; RCT, randomized control trial; Vim, ventral intermediate nucleus of the thalamus; Vom, ventralis oralis pars medialis.
Results
The literature search yielded 487 studies, of which 130 were duplicates (Fig. 1). A total of 357 studies were screened, of which 317 were deemed not relevant; of these, 36 were retrospective reviews and case series of pediatric DBS; most were mainly focused on dystonia or epilepsy. There were four studies that included Tourette’s syndrome, and one study on post-trauma tremors, and another that included cases of DBS done for GNAO-1-related movement disorders. The remaining studies omitted at the abstract and screening stage were for adult patients (e.g., Parkinson’s disease), non-DBS-related neuromonitoring (e.g., TMS), review papers, or anima/preclinical studies. With the addition of 2 studies identified via text references and a second literature search, 42 studies underwent full-text review. Thirty-three studies were excluded due to a predominantly adult patient population (N = 18), non-DBS intervention (N = 7), insufficient data (N = 6), nonclinical DBS outcome (N = 1), and nonclinical/prospective study design (N = 1). Overall, 9 published studies were included in this review, which collectively included 78 participants across 3 study designs (prospective study, N = 6, phase I clinical trial, N = 1; randomized control trial, N = 2), published between 2002 and 2024 (Fig. 2a). Among these studies, 6 focused on dystonia, and 3 on epilepsy. The country of study origin was diverse, including France (N = 2), Germany (N = 2), USA (N = 2), Canada (N = 1), Mexico (N = 1), and Spain (N = 1). Overall, 3 studies relied on industry funding [21, 30, 28].
Summary figures of the nine clinical trials included in this review (a) and the four ongoing clinical trials (b), organized by indication.
Summary figures of the nine clinical trials included in this review (a) and the four ongoing clinical trials (b), organized by indication.
Dystonia
The majority of published clinical trials studied dystonia (Table 1). Of the 6 studies, four had no specific inclusion/exclusion criteria [26‒28, 30], while the others varied in these criteria. The other two studies specifically recruited patients with acquired dystonia (i.e., exclusion of inherited forms of dystonia) [25, 29]. The goals and primary outcomes of each study also differed. For example, Candela et al. [27] and Starr et al. [30] focused on the safety and accuracy of robot- and MRI-guided DBS placement, respectively. Borggraefe et al. [26] determined whether the TOR1A mutation was associated with a positive therapeutic response to DBS. Furthermore, Legros et al. [28] specifically evaluated changes in resting tone and posture with DBS.
For 5/6 dystonia studies, the primary outcome was the Burke-Fahn Marsden Dystonia Rating Scale (BFMDRS). Overall, four studies reported a decrease in post-DBS BFMDRS scores (i.e., lower BFMDRS scores indicate less impairment), with an average improvement in dystonic symptoms of 1.6 ± 13.6–75.4 ± 12.5 across 6–12 months [26‒28, 30]. On the other hand, in San Luciano et al. [29], post-DBS BFMDRS scores after 12 months were higher than pre-DBS baselines (i.e., clinical worsening with bilateral Vom/Vim DBS for dystonia secondary to perinatal brain injury or encephalitis). However, clinical outcomes varied widely among participants, with differences between pre- and post-DBS BFMDRS scores ranging from +0.5 to −25 points, and no overall significant clinical difference was observed with stimulation. Koy et al. [25] used the Caregiver Priorities and Child Health Index of Life with Disabilities (CPCHILD) [34] as a primary outcome. CPCHILD approximates a patient’s overall health and functional status by relying on proxy measures such as caregiver burden, activities of daily living, communication, and mobility [34]. There was an insignificant 4-point increase in CPCHILD scores after DBS (i.e., higher CPCHILD scores indicate less impairment), with considerable variation between patients. However, there was a significant improvement in the Canadian Occupational Performance Measure (COPM, 1.1 ± 1.5 points [95% CI 0.2–1.9; p = 0.020]), a secondary outcome measure. COPM is an assessment tool to evaluate changes in the self-perception of productivity, self-care, and leisure. There was also a significant improvement in the physical health component of the 36-Item Short Form Survey (SF-36, 5.1 ± 6.2 points [95% CI 0.7–9.6; p = 0.028]), a widely used quality-of-life outcome measure.
The most commonly studied target for dystonia in pediatrics was the globus pallidus internus (GPi), known to be a well-established DBS target for dystonia in adult populations [35‒37]. Other targets included the bilateral ventralis oralis posterior/ventralis intermedius (Vop/Vim) [29], and 1 patient in Starr et al. [30] underwent subthalamic nucleus (STN)-DBS as they were enrolled in another trial for cervical dystonia [38].
We performed a meta-analysis of 5 of the 6 studies focusing on dystonia, which included individual patient-level BFMDS data allowing for pooled analysis. A BFMDS improvement of >16.6% was deemed to be clinically significant [39]. The pooled proportion of post-DBS BFMDS improvement was 0.91 (95% CI 0.77–0.99) (Fig. 3), with moderate heterogeneity across these (I2 = 20%).
Forest plot of fixed-effects proportional meta-analysis on proportion of patients with significant post-DBS BFMDRS improvement.
Forest plot of fixed-effects proportional meta-analysis on proportion of patients with significant post-DBS BFMDRS improvement.
Epilepsy
Three clinical trials were completed in epilepsy, two of which were small, single-arm prospective studies. Charbadès et al. [31] evaluated 5 patients after bilateral STN-DBS (two adult participants were omitted in the current analysis). There was no significant difference in seizure frequency after DBS in 1 pediatric patient, while the other 2 patients had a 67.8% and 80.7% reduction in monthly seizure frequency. The STN was selected based on early animal studies and the author’s previous success with STN-DBS for Parkinson’s disease. Alternatively, Velasco et al. [33] targeted the centromedian nucleus of the thalamus (CM) for Lennox-Gastaut syndrome in 9 patients. Bilateral CM-DBS was selected based on the author’s previous success with this target in a subset of adult patients with this childhood-onset epilepsy syndrome [40]. The average seizure reduction rate was 52.0% across 9 pediatric patients (range 0–100%), with 6 patients experiencing an overall reduction in seizure frequency [33]. A larger single-center study compared add-on DBS versus continued vagus nerve stimulation for childhood epilepsy (ADVANCE) [32]. This partially randomized patient preference comparative trial evaluated the role of add-on anterior nucleus or CM-DBS in children who had recurrent seizures after vagus nerve stimulation. At the 12-month follow-up visit, there was a significant reduction in seizure frequency with DBS (51.9% seizure reduction with VNS+DBS vs. 12.3% seizure reduction with VNS only, p = 0.047).
Risk of Bias Assessment
Bias risk was assessed using the ROBINS-I tool for each study (online suppl. Table 1). All published studies had a profound risk of bias, owing to the inherent limitations of single-arm studies relying solely on pretest/posttest measurements. Of the 7 domains comprising ROBINS-I, all studies had a serious risk of bias in at least one domain. This most commonly occurred due to deviation from the intended intervention, given the presumed systematic differences between the pretest and posttest individuals. Studies exempt from bias risk assessment were the crossover RCT by Koy et al. [25] and the study by Legros et al. [28], which provided comparisons to healthy controls. Suresh et al. [32] conducted a patient-selection RCT, which evaluated the role of add-on DBS for patients who did not respond to vagal nerve stimulation; there were a subset of 4-children that were initially in the control arm but crossed over to the DBS group. All but 3 studies [25, 27, 28] also had a serious risk of bias concerning participant selection due to poorly defined inclusion/exclusion criteria.
Adverse Events
Adverse events were documented in 5 of 9 studies, totaling 18 events across 63 participants. However, it is unclear whether these represent 18 unique adverse events or if participants had more than one adverse event. Documented events included infection (N = 4), transient neurological deficits (N = 3), device error (N = 2), and headache (N = 2), as well as isolated reports of device migration (N = 1), intracranial hemorrhage (N = 1), worsening symptoms (N = 1), fatigue (N = 1), hypersalivation (N = 1), skin seroma (N = 1), and postoperative pain (N = 1).
Ongoing Clinical Trials
A comprehensive search of the US-based trial registry, ClinicalTrials.gov, was also performed to assess ongoing, unpublished clinical trials of DBS in pediatrics. This search resulted in 4 trials currently investigating pediatric DBS for dystonia (N = 1) and epilepsy (N = 1), as well as novel indications, including cerebral palsy (N = 1) and self-injurious behavior (SIB) associated with autism spectrum disorder (N = 1) (Table 2; Fig. 2b). Among the ongoing clinical trials, 1 randomized control trial with a published protocol is examining the efficacy of cerebellar (dentate nucleus) DBS for dyskinetic cerebral palsy (NCT06122675) [41]. The other three trials consist of single-arm prospective studies, including the Children’s Adaptive DBS for Epilepsy Trial (CADET, NCT05437393), Thalamic DBS for Secondary Dystonia in Children and Young Adults (DBSVop, NCT03078816), and bilateral nucleus accumbens DBS in Children with Severe Self-Injurious Behaviours (NCT03982888).
Ongoing registered clinical trials of DBS for pediatric populations
Study . | Study design . | Location . | Diagnosis . | DBS target . | Outcome measure . |
---|---|---|---|---|---|
Children’s Adaptive Deep Brain Stimulation for Epilepsy Trial (CADET): Pilot (CADET Pilot, NCT05437393) | Single-arm, multi-site | London, UK | Epilepsy (Lennox-Gastaut syndrome) | CMN (bilateral) | Feasibility and safety trial |
Deep Brain Stimulation in Children with Severe Self-Injurious Behaviours (NCT03982888) | Phase I, nonblinded, nonrandomized, pilot trial | Toronto, Canada | Self-injurious behaviors in autism spectrum disorder | Nucleus accumbens (bilateral) | Changes in repetitive behaviors (Repetitive Behaviors Scale-Revised) |
Thalamic Deep Brain Stimulation for Secondary Dystonia in Children and Young Adults (DBSVop, NCT03078816) | Open-label single-group trial | San Francisco, USA | Dystonia | Thalamus (bilateral) | Change from baseline in Burke-Fahn-Marsden Dystonia Rating Scale |
Cerebellar Deep Brain Stimulation for Movement Disorders in Cerebral Palsy in Children and Young Adults (NCT06122675) | Randomized, crossover trial | San Francisco, USA | Cerebral palsy | Cerebellum (bilateral) | Movement Disorder-Childhood Rating Scale 4–18 Revised |
Study . | Study design . | Location . | Diagnosis . | DBS target . | Outcome measure . |
---|---|---|---|---|---|
Children’s Adaptive Deep Brain Stimulation for Epilepsy Trial (CADET): Pilot (CADET Pilot, NCT05437393) | Single-arm, multi-site | London, UK | Epilepsy (Lennox-Gastaut syndrome) | CMN (bilateral) | Feasibility and safety trial |
Deep Brain Stimulation in Children with Severe Self-Injurious Behaviours (NCT03982888) | Phase I, nonblinded, nonrandomized, pilot trial | Toronto, Canada | Self-injurious behaviors in autism spectrum disorder | Nucleus accumbens (bilateral) | Changes in repetitive behaviors (Repetitive Behaviors Scale-Revised) |
Thalamic Deep Brain Stimulation for Secondary Dystonia in Children and Young Adults (DBSVop, NCT03078816) | Open-label single-group trial | San Francisco, USA | Dystonia | Thalamus (bilateral) | Change from baseline in Burke-Fahn-Marsden Dystonia Rating Scale |
Cerebellar Deep Brain Stimulation for Movement Disorders in Cerebral Palsy in Children and Young Adults (NCT06122675) | Randomized, crossover trial | San Francisco, USA | Cerebral palsy | Cerebellum (bilateral) | Movement Disorder-Childhood Rating Scale 4–18 Revised |
ADVANCE, add-on deep brain stimulation versus continued vagus nerve stimulation for childhood epilepsy; CADET, Children’s Adaptive Deep Brain Stimulation for Epilepsy Trial; CMN, centromedian nucleus.
Discussion
Compared to adult populations, the literature pertaining to DBS in pediatric populations remains sparse. The current scoping review summarizes the landscape of completed and ongoing pediatric clinical trials, excluding less rigorous study designs, including case series. Compared to adults, where there are nearly 400 registered clinical trials, only 13 (9 completed, 4 ongoing) trials were identified in pediatric populations [9]. We describe limited, albeit expanding indications that include cerebral palsy and SIB, small sample sizes, and a high risk of bias. Further clinical research is needed to evaluate the role of DBS in children for a broader range of neurological and neurodevelopmental disorders.
The most commonly studied indication for pediatric DBS in the context of a clinical trial was dystonia, with 6 of 9 identified studies in this review focusing on this neurological disorder. Dystonia is a hyperkinetic movement disorder characterized by abnormal posturing, contractures, and repetitive motions. There are several axes of classification, including clinical features such as age of onset, anatomical distribution of symptoms, associated features, timing of symptoms, and inherited vs. acquired dystonia [42]. Based on the efficacy of GPi DBS for adult movement disorders, the Food and Drug Administration (FDA) approved a Humanitarian Device Exemption (HDE) for GPi DBS in children with dystonia aged 7 and older [43]. The meta-analysis of the 5 studies using BFMDRS as the primary outcome revealed a pooled proportion of improvement of 0.91; given the heterogeneity of the disease and outcome measurements, it is difficult to make direct comparisons to historical adult cohorts. Notably, due to the limited studies in our review, we were unable to group studies or patients based on the type/etiology of dystonia.
It may not be necessary to conduct a large randomized study for established indications such as GPi-DBS for generalized or segmental torsional dystonia. The positive open-label data show beyond equipoise that treatment may be beneficial and lessons gleaned from the treatment of adults may inform decisions in pediatric populations. There remains, however, much to be understood for the treatment of dystonia. For example, the investigation of other targets, including the Vom/Vim or dentate nucleus [41, 44], warrants further rigorous testing. Further trials for populations outside of the FDA HDE are also worthwhile, including DBS in children under the age of 7 years, different programming parameters, and studies to assess other quality-of-life-based measures particularly in the context of acquired dystonia [19, 20].
Ongoing research into biomarkers and relevant outcome measures of response to DBS is also important, particularly given that the optimal outcome for clinical trials of DBS in children remains elusive. In dystonia, the BFMDRS remains a widely used rating score in clinical practice and research. Despite its ubiquity, a major limitation of this scale is the uncertainty regarding clinically significant changes. Historically, studies use an arbitrary 25% cut-off for clinical efficacy, while Pintér et al. [45] demonstrated that any improvement greater than 16.6% was clinically significant. Another consideration is that BFMDRS improvement may not correlate with the quality of life improvement [46], which is better captured on scales such as the PedsQL™ Measurement Model [47], which is not a standard metric used in the DBS literature [48]. Some newer measures, such as the Caregiver Priorities and Child Health Index of Life with Disabilities [34] questionnaire used by Koy et al. [25], aim to characterize overall quality of life and well-being by accounting for activities of daily life, mobility, and emotional changes. This may capture clinical changes more meaningfully for the individual patient, particularly children, and their caregivers.
Epilepsy was the second most common indication for pediatric DBS, with 30 patients implanted. Overall, the mean reduction in seizure frequency across the participants included in this review was 52.0%, compared to a 75% mean reduction at 7 years in adult patients [49], possibly indicating unique outcomes in pediatric cohorts. Only 1 of 3 epilepsy studies targeted the anterior thalamic nucleus, the most widely used target for drug-resistant epilepsy in adults, likely reflecting differences in epilepsy syndromes between adults and children [50]. Charbadès et al. [31] targeted the STN, a less common target for epilepsy but thought to be well suited for managing motor/myoclonic seizures [50]. The centromedian nucleus (CMN) was used as a target by Velasco et al. [33] for patients with intractable Lennox-Gastaut syndrome, based on the results of an earlier study in adults [40]. One of the ongoing clinical trials (CADET, NCT05437393) is also studying bilateral CMN-DBS for children with Lennox-Gastaut syndrome but using adaptive technologies. These and future studies will clarify the optimal DBS target and stimulation paradigms for refractory pediatric epilepsy.
One alternative explanation for the dearth of clinical research in DBS for novel indications in children is that DBS is not understudied but rather may not be necessary, useful, or pragmatic. Such a view is contradicted by an emerging body of preclinical research in childhood disorders that has highlighted common circuitopathies that may be amenable to neuromodulation. For example, disordered impulsivity is common to numerous childhood-onset neurodevelopmental disorders including attention deficit/hyperactivity disorder [51, 52] and severe SIB [53, 54] in children with autism spectrum disorder. Rodent models of repetitive, stereotyped, and compulsive behavior demonstrate hyperactivity in the striatal interneurons and their cortical projections [55, 56]. Optogenetic manipulation of these striatal projections modulates the frequency and severity of excessive grooming [57, 58], the rodent equivalent of SIB [59]. These preclinical studies have motivated an ongoing phase I study in children (NCT03982888).
In this study, half (i.e., 5 of 9) studies reported adverse events, most of which were hardware-related (i.e., 8 of 18 events). This is in keeping with previous reports of more hardware-related complications in pediatric populations (up to 18% [60]) compared to adults (approximately 8% in adult counterparts [61]). This may relate to the population itself, the different etiologies being treatment or the device. This DBS leads and implantable pulse generator are designed for adults, children have a smaller proportion of subcutaneous fat in children, and there is a more significant cumulative lifetime risk of infection with repeated need for batter replacements. Some of these risks may be mitigated through the development of smaller, more flexible, or antibiotic impregnated devices [62].
This review also highlighted the small number of active clinical trials dedicated to pediatric DBS (N = 4) compared to nearly 400 trials for adult patients across multiple indications and targets [9]. While there is a clear need for well-designed studies based on valid neuroscience in pediatric populations, the enthusiasms for such studies need to be tempered by pragmatism and ethical considerations. Children represent an exquisitely vulnerable patient population, and research efforts need to recognize this unique status. There is a need for clinical equipoise prior to engaging in clinical research, which may not exist in the presence of well-conducted retrospective studies or in the setting of translational data from the adult literature. Furthermore, we have previously published an ethical framework for the conduct of DBS clinical research in pediatric populations [22]. The five tenants of the framework include: (1) study design and conduct in the context of the best interest of the child; (2) hypotheses, outcomes, and procedures viewed through a developmental lens; (3) discussion of what is known and unknown regarding risk profile both, including timing of intervention; (4) cautious application of adult data to pediatrics; and (5) diligent reporting of positive and negative studies.
We also highlight several limitations to the current scoping review. Although the primary purpose of the current scoping review was to highlight gaps in knowledge to inform future directions within the field of pediatric DBS, we did attempt a meta-analysis of included studies. We were only able to conduct this meta-analysis of five studies reporting outcomes using BFMDRS. There was moderate heterogeneity among these included studies, limiting the generalizability of these findings. Second, pediatric specific considerations were often underreported, including head fixation techniques, target accuracy assessments, and use of microelectrode recordings. Finally, we did not include mixed studies of pediatric and adult cohorts. Most of these studies were in older literature and did not provide individual participant data. The current work can therefore be understood as an overview of knowledge gaps in pediatric-specific research.
Conclusions
This scoping review provides an overview of the clinical trial landscape for pediatric DBS. Dystonia and epilepsy were the most commonly studied indications, with generally positive outcomes reported. The literature in pediatric populations in general remains sparse. While DBS remains a promising therapy for refractory neurological and psychiatric disorders in adults, it has not been studied in children. Larger, well-designed studies with rigorous methodology and biologically sound target selection and stimulation parameters are needed to clarify the efficacy and scope of pediatric DBS.
Statement of Ethics
A statement of ethics is not applicable because this study is based exclusively on the published literature.
Conflict of Interest Statement
G.M.I. is an advisor for LivaNova Inc., Medtronic Inc., and Synergia Inc. C.G. is an advisor for Medtronic Inc. and Ipsen Inc. A.G.W. is a consultant for Monteris Medical Inc. All other authors declare no competing interests.
Funding Sources
This work was supported by Abe Bresver Chair in Functional Neurosurgery (G.M.I.), the Hospital for Sick Children Centre for Brain and Mental Health Program Development Grant (G.M.I., C.G.), the Canadian Institutes of Health Research (I.E.H.), and the UofT/UHN Chair in Neuromodulation (A.F.). The authors have no financial supports or sources of funding to disclose.
Author Contributions
Y.J., K.M., and I.E.H.: conceptualization, data curation, writing – original draft, and writing – review and editing; H.S.: data curation and writing – review and editing; N.W. and S.B.: conceptualization and writing – review and editing; and C.G., A.F. (Fasano), A.H., A.W., and G.M.I.: conceptualization, writing – original draft, and writing – review and editing. All authors reviewed and approved the final version of the manuscript.
Data Availability Statement
This paper is a systematic review; the data are all derived from previously published trials and papers. The databases we used to search for our primary articles are outlined in the Methods section and in our supplementary materials. All further inquiries may be directed towards the corresponding author.