As pharmacological management is dissatisfying in many metabolic movement disorders, neurosurgical procedures have been attempted to alleviate the most disabling symptoms, which are often severe dystonia, chorea, tremor, and self-mutilating behavior. These neurosurgical procedures include ablative procedures and deep brain stimulation (DBS), typically targeting the basal ganglia.
As pharmacological management is dissatisfying in many metabolic movement disorders, neurosurgical procedures have been attempted to alleviate the most disabling symptoms, which are often severe dystonia, chorea, tremor, and self-mutilating behavior. These neurosurgical procedures include ablative procedures and deep brain stimulation (DBS), typically targeting the basal ganglia. DBS is a neuromodulation technique in which electrodes are precisely implanted in deep brain structures, such as the basal ganglia, by applying methods based on brain imaging and coordinate systems. Electrical pulses are then delivered by an implantable pulse generator, thereby modulating the neural tissue in the vicinity of the electrodes. Although the underlying mechanism is not completely understood, the effect of DBS is usually very similar to that of ablative procedures. However, DBS has the advantage of being adaptable and (at least partially) reversible.
Unlike many other indications for DBS, large randomized blinded crossover trials, directly comparing DBS on vs. off, are not available for metabolic movement disorders. Conducting such trials is difficult for various reasons. First, most metabolic movement disorders are very rare and large trials would face slow recruitment. Moreover, these disorders are often heterogeneous in terms of genotype and phenotype . Next, practical difficulties could arise from long wash-in and wash-out periods as reported in DBS for many forms of dystonia [2, 3]. Further, direct DBS on vs. off comparisons may be deemed unethical, because of the risk of life-threatening rebound effects such as dystonic storms  and the psychological issues associated with the cessation of DBS only for scientific purposes . Lastly, many patients are not able to provide consent because they are too young or have intellectual disability and associated communication difficulties. Non-randomized studies including an open control group of patients who were offered but refused to undergo DBS are scientifically less valid, but may represent a more pragmatic alternative .
The scientific evidence for DBS in metabolic movement disorders is therefore limited to case reports or small case series, with a high likelihood of publication bias, causing under-reporting of negative outcomes. In the majority of publications, neither patients nor assessors were blinded to the procedure, opening the possibility of the placebo effect or observer bias. Moreover, DBS on vs. off comparisons are almost completely lacking. Consequently, the effect of DBS on the natural course of the disease and vice versa is unclear, and the additional benefit of DBS over the sometimes long-lasting insertional effect has not been demonstrated [7–9].
General Considerations in DBS for Metabolic Movement Disorders
Any treatment for movement disorders should aim to improve functionality and quality of life by improving movement and posture and relieving any associated disability, pain, and discomfort . Assessing symptom severity with a uniform scale is difficult as metabolic movement disorders vary largely in age at presentation and symptoms. Table 31.1 provides an overview of the frequently used scales in metabolic movement disorders.
As dystonia is the main indication for DBS in metabolic movement disorders, most publications report the Burke–Fahn–Marsden Dystonia Rating Scale (BFMDRS) or Barry–Albright Dystonia Scale (BADS). Chorea/ballism and other involuntary movements are generally rated through the Unified Huntington Disease Rating Scale (UHDRS) or Abnormal Involuntary Movement Scale (AIMS). However, as BFMDRS was originally established for the quantitative assessment of primary torsion dystonia in adults, it is limited to the assessment of a single symptom of a movement disorder and does not discriminate between abnormal postures and movements caused by non-dystonic symptoms. Moreover, scales such as the BFMDRS and AIMS are relatively insensitive to clinically meaningful changes in axial dystonia, vocalization, communication, fine motor skills, hand dexterity, and swallowing . Even the authors of the BFMDRS-D acknowledge that “major changes are required before the change registers on the scale” .
Many publications on DBS for metabolic movement disorders emphasize that the reported scales do not fully express the subjective benefit experienced by the patient, the caregiver, or the physician because they are not fine enough or because they do not take into account certain activities which can be highly relevant to the quality of life . In the given context, what may seem a minor functional improvement can be of utmost importance to the patient or caregivers. Although this underscores the need for disease- and age-specific severity scales, the main function of these scales remains to objectively compare symptoms and their impact, within but also between patients. Therefore, over-refinement of scales to perfectly match a small subset of patients is not beneficial per se. Moreover, when seeking approval and reimbursement for DBS in these disorders, a direct comparison with other treatment options, such as medications, requires the continued use of relatively generic scales.
In metabolic movement disorders in general, patients who have isolated dystonia, chorea/ballism, or tremor (no other combined symptoms), without fixed skeletal deformities or muscle contractures, and who have experienced a short duration of symptoms, are probably the best candidates for DBS . Therefore, discussing the possibility of DBS with the patient and his or her caregivers is justified as soon as the functional capacity and/or quality of life are impaired despite optimal medical management, or when the side effects of medical treatment are intolerable. Brain imaging is paramount to assess the status of the intended target, although surgery is not necessarily precluded if the target is severely compromised (e.g. pallidal stimulation in pantothenate kinase-associated neurodegeneration [PKAN]). Nevertheless, patients with a long-standing disease, who already display skeletal deformities and contractures, can still benefit from DBS, particularly in PKAN. Importantly, DBS is not only intended to improve the quality of life but can also become an emergent last resort in patients presenting with intractable dystonic storm.
The best results can probably be obtained when the patient and caregivers are assessed, counseled, operated, programmed, and followed by an experienced and interdisciplinary movement disorder team, consisting not only of (pediatric) neurologists and neurosurgeons, but also psychiatrists, neuropsychologists, and physiotherapists.
Many patients with metabolic movement disorders are children, especially those with PKAN or Lesch–Nyhan disease, and hence special considerations apply.
General anesthesia is usually required in these children with severe movement disorders, not only for the surgery itself, but also for the frame placement and preoperative imaging [14–16]. Electrophysiological measurements to confirm the electrode position are therefore omitted  or are performed under anesthetic conditions that make some recordings possible, often under bispectral monitoring . Direct anatomical targeting and accuracy verification using intraoperative brain imaging can be proposed when electrophysiological measurements are impossible .
For the neurosurgeon, it is important to take care when adjusting the pressure applied to the frame pins to the softer and thinner skull in children, especially in those with metabolic diseases such as abetalipoproteinemia. In children ≤7 years of age, implantation of the DBS electrode so that its deepest contact is ventral to the target area is reasonable to prevent the loss of effect due to the growth of the brain and the skull [20, 21]. While tunneling, enough slack should be inserted in the extension wires to avoid a hardware fracture because of body growth, and in girls, the breast buds should be avoided . As body weight is often low, especially in dystonic patients and/or those with dysphagia, the implantable pulse generator (IPG) is often implanted in an abdominal wall pocket rather than subclavicularly . Other technical pearls sometimes include staged procedures or the prolonged use of head wraps to prevent cerebrospinal fluid (CSF) leakage along the wires to the IPG pocket, suturing the connector to the nuchal fascia, or burying the connector in a partial skull thickness groove and using two completely separate systems in case of bilateral stimulation (although there is a risk of the IPGs interfering with each other when positioned too closely together, which can be a consideration in the pediatric population) [15, 22]. Proposals to reduce the infection risk of DBS in children have been described .
However, despite all these measures, the postoperative risk of infections and hardware complications in children remains substantial, probably because of their immature immune system, limited tissue to cover the devices, and stress induced by growth or physical activities [22, 25, 26].
From a medicolegal point of view, it is noteworthy that commonly used DBS systems are not approved in young children (e.g. the Medtronic Activa system is only approved by the US Food and Drug Administration in children age 7 years and older) . Their use in certain situations and/or jurisdictions may warrant single-use institutional review board approval.
The general anesthetic rule in patients suffering metabolic diseases is to avoid perioperative metabolic decompensation, which may be triggered by prolonged preoperative starvation, dehydration, hypoxia, hypotension, hypothermia, stress response during surgery, and certain anesthetic agents. Adequate perioperative analgesia reduces the stress response. In general, patients with metabolic disorders are prone to develop osteoporosis, and therefore special attention to frame placement, surgical positioning, and transfer is required [15, 21]. Anesthetics should be administered by a team familiar with the nuances of agent selection in the context of metabolic disorders.
Typical DBS targets for movement disorders are the posteroventral globus pallidus internus (GPi), the subthalamic nucleus (STN), and the motor thalamus (ventral intermediate [Vim] and ventral oral posterior [Vop] nuclei)). All of these targets are within a neural circuit connecting the motor cortex, basal ganglia, and brainstem. In metabolic movement disorders, thus far only these three targets have been reported, with most patients undergoing bilateral GPi-DBS.
The most frequent and deleterious complications of DBS in general include stroke, infection, skin erosion, electrode malposition, and hardware failure or breakage. In metabolic movement disorders, it appears that the infection rate is low, even in young children. Nevertheless, in children, there is a relatively high incidence of hardware problems with the electrodes and extension wires . Electrode malposition is an infrequent problem, irrespective of the degree of basal ganglia degeneration .
It often takes several weeks of stimulation before the tonic component of dystonia starts to improve, and even several months of stimulation before this effect is maximal. Therefore, sufficient patience is paramount. Moreover, maximizing the benefit and reducing the side effects may take many reprogramming sessions, again taking into account the protracted time to potentially realize benefits. Importantly, DBS should be considered as an addition to ongoing treatment, not as a replacement. Therefore, continuation of medical treatment with necessary adaptations, and follow-up and extensive rehabilitation, are essential for patients who lose functionality in the period immediately preceding DBS surgery. The acute cessation of stimulation, typically because of battery depletion, hardware fracture, or system removal because of infection, has the potential to elicit a life-threatening status dystonicus. When suspected, battery replacement or revision surgery should be performed emergently . In the case of system removal because of infection, radiofrequency ablation through the DBS system preceding removal can be considered.
Performing DBS in childhood, in adults with impaired cognition and/or communication, or during a life-threatening dystonic storm in an anesthetized patient, requires careful ethical consideration. Moreover, the parents of these vulnerable patients struggle in the face of uncertainty over the outcome whilst trying to do their best as parents, and therefore require adequate social support .
The largest experience with DBS for metabolic movement disorders exists in PKAN, choreo-acanthocytosis, Lesch–Nyhan disease, kernicterus, and glutaric aciduria type 1. Other metabolic movement disorders, with less than five DBS cases reported in the literature, will be discussed only briefly.
PKAN is a rare (prevalence 1–2 in 1,000,000) autosomal-recessive disorder. It is characterized by progressive iron accumulation preferentially affecting the globus pallidus and pars reticularis of the substantia nigra, typically as a result of PANK2 gene mutations [30, 31]. PANK2 encodes an enzyme of the same name, which has an important role in coenzyme A synthesis. Coenzyme A plays a central role in fatty acid synthesis and energy metabolism. Coenzyme A deficiency can lead to a high concentration of cysteine in the basal ganglia, and then to iron accumulation. By provoking oxidative stress, the cysteine–iron complex results in tissue damage .
PKAN is associated with progressive extrapyramidal findings (generalized dystonia, rigidity, choreoathetosis), corticospinal tract dysfunction, and cognitive deterioration. Typically, there is a so-called “eye of the tiger” sign on T2-weighted MRI, referring to a hypointensity with central hyperintensity in the globus pallidus . The hypointensity corresponds to iron deposits in dense tissue, and the hyperintensity appears pathologically as an area of loose tissue with vacuolization . Many publications have made a distinction between two types of PKAN, based on the age of onset and rate of progression. Classic PKAN is characterized by early-childhood-onset of progressive dystonia, dysarthria, rigidity, and choreoathetosis. Pigmentary retinal degeneration and optic atrophy are common. Atypical PKAN is characterized by later onset (typically >15 years of age), prominent speech defects, psychiatric disturbances, and a more gradual progression of disease . Pharmacological therapy with dopamine, clonazepam, trihexyphenidyl, tetrabenazine, and baclofen is generally unsatisfactory.
Medication-refractory cases with severe functional impairment are candidates for neurosurgical treatment. Neurosurgical options include intrathecal baclofen delivery using pumps , pallidotomy , thalamotomy , pallidothalamotomy , and DBS targeting the GPi, or alternatively the STN or motor thalamus. Our systematic review of the literature encompasses 99 cases of DBS for PKAN (58 classic type, 15 atypical type, 27 undetermined type), published since 2001 . GPi is the most common target (88%). It is intriguing that, although severely affected by iron accumulation, neuronal degeneration, and gliosis, the GPi still contains viable neurons (as demonstrated during microelectrode recording [MER]) and that these neurons can be stimulated with a clear benefit . Although clearly efficacious for Parkinson dystonia, the benefit of STN-DBS for PKAN dystonia is highly variable, but may be an option for appendicular dystonia [42, 43]. In general, thalamic DBS appears inferior to GPi-DBS for dystonia .
The ideal timing of DBS in the course of the disease has not yet been determined. Patients with classic PKAN underwent DBS implantation surgery at the median age of 11 years (range 4–36 years) after a median disease duration of 7 years. Atypical PKAN cases treated with DBS are operated on at a median age of 31 years (range 17–46 years), after a median disease duration of 15 years. Ideally, the surgical procedure should be performed electively to allow sufficient time for counseling and preparation. However, urgent DBS surgery for medically refractory status dystonicus is not uncommon in classic PKAN, with 10% of published cases operated in status dystonicus .
Essentially, the surgical technique of DBS for PKAN is the same as in DBS for any type of dystonia. However, there are at least three nuances. First, many classic PKAN cases are operated during childhood (as early as 6 years of age). In these cases, all considerations for pediatric DBS (see above) apply. Second, there are many anesthetic challenges, before and after anesthesia (e.g. articulation difficulties and cognitive impairment limiting communication), during induction (e.g. oromandibular rigidity, dysphagia, aspiration), during anesthesia (e.g. seizures), and postoperatively (e.g. respiratory disability) . Third, whether or not MER should be used, and if so, under local or general anesthesia, remains a matter of debate. Conflicting MER characteristics of GPi neurons have been reported in PKAN patients. Shields et al. published quiet areas and areas that show a 37.5-Hz irregular firing pattern in an awake patient , McClelland et al. found a regular 25-Hz firing rate in eight patients under isoflurane anesthesia , similar to Valentin et al. who found a discharge rate of 35 Hz under sevo/isoflurane anesthesia in six patients , while Justesen et al. reported a regular firing rate of 91 Hz in a single patient under light propofol anesthesia . Despite these variable firing rates, MER can certainly be useful to localize the optic tract and hence the ventral GPi border. The MER is, however, not an absolute requirement and there are variable techniques utilized between different specialized centers, with some using purely radiological-based approaches to place the DBS electrodes.
Programming is usually initiated between 1 day  and several weeks  after electrode implantation, although in cases of dystonic storm immediate programming is reasonable. As for other types of dystonia, stimulation settings vary largely, and there is no general agreement on the best combination or programming algorithm . Typically, amplitudes in GPi-DBS increase slowly over time  from 1 V to 5 V (median 2.0 V), with a frequency ranging from 120 Hz to 185 Hz (median 130 Hz) in a monopolar or double monopolar configuration. The median programmed pulse width is 140 μs–145 μs, although some groups use much larger pulse widths (450 μs) [40, 50].
Given the risk of a dystonic storm after battery depletion, planning follow-up visits and/or instructing caregivers to assess the battery status are paramount. If the battery depletes unexpectedly, close monitoring and semi-urgent replacement must be considered based on the patient’s clinical status.
DBS for PKAN has been reported to result in a considerable reduction of dystonia severity, similar to [44, 50] or slightly smaller  than in primary dystonia (DYT-1). However, the outcome in PKAN is much more heterogeneous. In our review of 99 published cases, the mean BFMDRS-M reduction after 1 year of –26% was more prominent in atypical (–45%) than in classic (–16%) PKAN cases. Patients with atypical PKAN also had a lower mean baseline BFMDRS-M than those with classic PKAN (52 vs. 82) . Many factors may contribute to the better outcome in atypical PKAN cases, including less fixed skeletal deformities and muscle contractures than in classic PKAN (0/15 vs. 8/58) and later disease onset (median of 15 vs. 6 years of age), thereby acquiring a higher level of motor skills and potentially facilitating postoperative rehabilitation.
As dystonic postures and abnormal movements improve substantially, medication doses can often be reduced [23, 52]. Walking capacity improves in the majority of patients, even those who were wheelchair-bound preoperatively . Speech and writing skills improve in ≥50% of cases . Oromandibular dystonia appears to respond less to GPi-DBS [41, 52]. Several publications report a major improvement of pain secondary to hypertonic muscle cramps, often resulting in a complete cessation of analgesics [50, 53]. Similar to dystonia severity, improvement in disability is more prominent in atypical PKAN than in classic PKAN (BFMDRS-D –31% vs. –1%), with a lower mean baseline disability in atypical vs. classic PKAN (BFMDRS-D 17 vs. 25) . Data on the evolution in the quality of life after DBS for PKAN is scarce. In a retrospective assessment on a 0–10 scale by the caregivers, an 80% improvement was noted 3 months postoperatively and sustained at 1 year. However, these numbers are likely affected by recall bias .
It appears that GPi-DBS for PKAN does not only improve dystonic features, pain, and quality of life, but also psychosocial wellbeing [3, 54] and cognitive test scores [52, 54, 55]. Whether the latter is due to reduced pain and hence reduced analgesics [50, 53], improved dystonia and hence reduced anticholinergic medication , improved speech  and hence easier responding , or true improvement in cognition and memory (e.g. via decreased response inhibition ) is difficult to determine .
The median reported follow-up is 1 year, and publications with a longer follow-up duration are scarce (7/99 ≥ 2 years and 2/99 ≥ 5 years). Conclusions drawn from these few patients are likely to be affected by selection bias, as patients with poor outcomes may not be followed as closely. Nevertheless, even in these selected cases, a gradual decline in efficacy is typically observed, with dystonia severity returning to baseline between 4 years and 7 years post implantation, despite extensive reprogramming and excluding hardware failure. One of the possible explanations for this phenomenon could be progressive pallidal neuronal degeneration, thereby confining the DBS substrate . As long-term DBS on vs. off comparisons are lacking, it is unclear whether DBS continues to improve dystonia in the long term. One year postoperatively, 27/56 (48%) and 17/56 (57%) of the PKAN patients experienced a ≥30% or ≥50% improvement in BFMDRS-M, respectively. Again, patients with atypical PKAN display a higher response rate (73% of atypical PKAN patients demonstrate >30% improvement vs. 35% of classic PKAN patients) .
In our analysis of 99 published cases, the preoperative dystonia severity, disease duration, and proportion of life lived with symptoms were not predictive of the relative BFMDRS-M reduction. Interestingly, in classic PKAN cases, there was no difference in outcome (relative BFMDRS-M reduction) between patients with and without fixed skeletal deformities and/or muscle contractures preoperatively. The numbers are, however, small and potentially affected by reporting bias .
The exact mechanism of the action of GPi-DBS in PKAN remains unknown. A technetium single-photon emission CT study demonstrated prominent tracer accumulation in the bilateral pallidum, which was completely reversed 9 weeks after successful GPi-DBS, although an artifact induced by insertional edema cannot be ruled out completely . An EEG study demonstrated the correction of electrophysiological responses during response inhibition tasks with GPi-DBS on vs. off .
The complication risk is not significantly different in children and adults with PKAN compared with what has been reported in other types of dystonia . In our review, 5/10 patients operated in status dystonicus had a complication, compared to only 1/18 patients who were explicitly reported not to be operated in status dystonicus. In 6/99 cases, an infection was reported, of which led to a brain abscess. Hardware problems were relatively common (three cable fractures causing status dystonicus in one patient; two IPG malfunctions; two lead migrations). In one patient, revision surgery was necessary because of inadequate targeting. Death was reported in nine patients, mostly because of end-stage disease >1 year postoperatively, although two patients died within 3 months. Another patient needed a tracheostomy 1 month postoperatively after an aspiration pneumonia. Status dystonicus resulting from surgery and anesthesia caused a spontaneous femur fracture immediately postoperatively in one patient and a hip subluxation in another. Stimulation-caused side effects included blepharospasm and worsening of gait with GPi-DBS and reduced verbal fluency and worsening of dystonia with STN-DBS. In the case of a hardware infection necessitating system removal, a pallidotomy through the infected electrode lead before removal can be considered to reduce the risk of status dystonicus.
There is Oxford Centre for Evidence-Based Medicine (OCEBM) level 4 evidence that in about half of patients undergoing surgery, GPi-DBS substantially reduces dystonia severity (mean BFMDRS-M reduction of –26%) and improves functionality, pain, cognitive scores, and quality of life at 1 year postoperatively. Patients with atypical PKAN benefit significantly more from DBS than those with classic PKAN. However, in the long term, this benefit is lost, although the relative contribution of disease progression and loss of DBS effect is unknown. The risk of infection and hardware malfunction is in line with other types of dystonia, although two patients died within 3 months and the complication rate was particularly high in patients operated in status dystonicus.
Choreoacanthocytosis (ChAc) is an autosomal-recessive disease with mutations in the VPS13A gene, encoding the protein chorein, for which the function is not well understood. ChAC is characterized by progressive cortical and basal ganglia neurodegeneration and abnormal red blood cell morphology (acanthocytes). It manifests generally in the third to fourth decade with a mixed movement disorder, predominantly in the orofacial region. Symptoms include dysarthria, chorea, hypotonia, and truncal spasms which often cause involuntary head banging. These are typically accompanied by cognitive decline, psychiatric manifestations, epilepsy, myopathy, and axonal neuropathy [60–62].
Many of these features resemble Huntington disease (HD) and ChAC symptom severity is often expressed using the UHDRS. However, ChAC patients generally progress more slowly than HD patients, and exhibit tongue- and lip-biting, self-mutilating behavior (SMB), and seizures, and these symptoms are usually not seen in HD .
ChAC is characterized by neuronal loss in the striatum, resulting in striatal atrophy and mild ventriculomegaly . The diagnosis can be confirmed by the western blot , or by sequencing of the VPS13A gene . A blood smear presence of acanthocytes is suggestive but not pathognomonic for ChAC.
The most debilitating symptoms, chorea and dystonia, are often poorly controlled by pharmacological therapy, which includes botulinum toxin injections, tetrabenazine, and atypical neuroleptics.
So far, DBS has been reported in 22 ChAC patients (17 male). The median age at surgery was 38.5 years (range 30–54 years) after a median disease duration of 7.5 years (range 1–24 years). In a multicenter retrospective case series of 15 patients, the indications for surgery were disabling motor symptoms (chorea, dystonia, trunk spasms, falls, and gait impairment; n = 11), SMB (n = 6), head banging induced by trunk spasms or head drops (n = 4), feeding dystonia (n = 3), and recurrent belching (n = 2) . None of the patients suffered a dystonic storm.
Following a report on a successful pallidotomy in ChAC , GPi-DBS for ChAC was first reported in 2001  and the GPi has been targeted bilaterally in all cases, except for one patient treated successfully with bilateral thalamic (Vo) stimulation . Nakano et al. also implanted Vo electrodes in addition to GPi electrodes. Interestingly, in these patients, GPi stimulation alone improved falls and oromandibular dystonia but was insufficient to reduce trunk spasm and chorea (19–27% UHDRS reduction 1 year postoperatively). Vo stimulation alone improved UHDRS more than GPi stimulation alone (33–38% UHDRS reduction), but combined GPi and Vo stimulation was superior (40–46% UHDRS reduction) .
From an anesthetic perspective, it seems that dexmedetomidine is a useful agent to reduce the orofacial dystonia, which may interfere with stereotactic MRI and/or the surgical procedure . The use of MER was reported in 6/22 patients, but no details about the characteristics of GPi neurons in ChAC patients have been published to date.
It usually takes approximately 1 month before improvement in oromandibular dyskinesias and limb chorea is observed, and it takes several months before the maximal benefit is obtained [62, 69, 70]. As in other types of dystonia, there is no consensus on the ideal DBS settings. Typically, the complexity increases when several types of abnormal movements are present simultaneously, such as a combination of chorea and dystonia, which is usually the case in ChAC. In a multicenter retrospective case series of 15 ChAC patients, the majority of patients required high-frequency stimulation (130–185 Hz), while in 4/14 patients, high-frequency stimulation worsened chorea and/or dystonia, and therefore low-frequency stimulation (40–60 Hz) was (initially) applied . Low-frequency stimulation may worsen truncal spasms . Very low-frequency stimulation (10 Hz) had no effect, and very high-frequency stimulation (>500 Hz) worsened the symptoms . Stimulation settings thus need to be highly individualized based on the presence of various symptoms and their relative contribution to global function . With longer follow-up, the amplitude, pulse width, and number of active contacts had to be gradually increased to maintain the stimulation benefit .
A multicenter retrospective case series of 15 patients reported on outcomes and covariates associated with better results for bilateral GPi-DBS in ChAC . On average, choreatic movements were substantially reduced early postoperatively, by 50% on the UHDRS-Movement Scale (UHDRS-MS)  and 38% on the AIMS [7, 69]). It appears that chorea (within hours), dystonia (within hours to days for the phasic component; within weeks to months for the tonic component), trunk spasms, head drops, orofacial movements, and SMB have a >90% chance of improvement with DBS. Dysarthria rarely improves . Gait improved in patients with chorea and dystonia, but not in patients with parkinsonism. Feeding status improved in 6/9 patients. In two patients severe truncal hypotonia did not respond to DBS [61, 73]. DBS did not influence the Mini-Mental Status Exam [72, 74]. Postoperative weight gain, most likely because of less impaired feeding, has been reported . Although no formal DBS on vs. off comparisons have been published, a rapid loss and resolution of effect after system removal and re-implantation, respectively, has been described . Moreover, EMG registrations have shown a more than threefold decrease in muscle spasms with DBS . The functional status, as measured by the UHDRS independence score (UHDRS-IS) and functional capacity score (UHDRS-FCS), improved shortly after surgery from on average 53 to 73 (+37%) and 4.1 to 7.2 (+76%), respectively. Seventy to 80% of patients had a clinically relevant functional improvement. In the long term, this functional improvement seems more stable than that of the UHDRS-MS . No formal quality of life evaluations before and after DBS have been published in ChAC.
In the long term, the UHDRS-MS reduction is not retained in all patients. At 3 and 5 years postoperatively, only 3/7 and 0/2 patients still had a ≥20% UHDRS-MS improvement compared to immediately preoperatively. As no long-term DBS on vs. off comparisons have been published, it cannot be determined with certainty whether the loss of effect is due to loss of DBS efficacy or due to disease progression, although the latter seems more plausible. Only 1/22 patients failed to respond to DBS (<20% improvement), but in this particular patient the electrodes were removed within 1 month postoperatively, after continuously changing the settings. It is possible that insufficient time for effect wash-in was permitted . The average UHDRS-MS reduction of approximately 50% is independent of the preoperative UHDRS-MS. There was no significant relationship between UHDRS-MS reduction and age or disease duration . The majority of patients were able to taper their medication following DBS (total cessation of medication in 1/14) [60, 61].
Little is known about the mechanism behind DBS for ChAC. Bilateral striatal hypometabolism, as detected by fluoro-2-deoxyglucose positron emission tomography in ChAC patients, remained unchanged with DBS . Seizures have been reported in one patient during DBS surgery (associated with a subdural hematoma) and in one patient 6 months after surgery. Given the predisposition for seizures in ChAC patients, the additional seizure risk caused by DBS appears to be low. Furthermore, one hardware infection resulting in a pallidotomy-like lesion, one lead pulled during an external trial, one Twiddler syndrome (i.e. malfunction due to manipulation of the device), and one sudden death 2 years post DBS implantation have been reported to date .
There is OCEBM level 4 evidence that GPi-DBS can improve chorea (UHDRS-MS, –50%; AIMS, –38%), dystonia, SMB, and functional status (UHDRS-IS, +37%; UHDRS-FCS, +76%) in ChAC patients. In the literature, >90% of cases responded to DBS. Medication can often be reduced. However, the improvement in chorea and dystonia vs. preoperatively is lost over approximately 5 years, most likely because of disease progression. Both high- and low-frequency stimulation can be beneficial for different motor components and therefore stimulation settings need to be individualized. The complication rate, including the risk of seizures, is low.
Lesch–Nyhan disease (LND) is an X-linked recessive disorder associated with mutations in the HPRT1 gene. This results in a complete deficiency of hypoxanthine–guanine phosphoribosyltransferase (HGPRT), which is a purine salvage enzyme [75, 76]. Clinically, LND expresses during early childhood with hyperuricemia leading to gout and nephrolithiasis, hypotonia evolving towards generalized dystonia, choreoathetosis, weakness and spasticity mainly in the lower limbs, and cognitive impairment. However, the most disabling symptoms are aggression and SMB, typically lip-biting or finger-chewing as soon as teeth are present [77, 78]. LND patients generally survive with a limited quality of life. Imaging studies  and neuropathological examinations  have not disclosed any abnormality except for striatal atrophy. Neurochemical analysis of post-mortem tissue and CSF revealed decreased levels of dopamine and its metabolite homovanillic acid, while serotonin and 5-hydroxyindoleacetic levels vary. Abnormal monoaminergic transmission in the basal ganglia is likely to play a key role in the pathophysiology [81, 82].
Treatment of LND involves limiting hyperuricemia to avoid nephropathy and gout, but normalization of uric acid levels does not alter the neurological symptoms of the disease. Hypertonia is frequently treated with diazepam, baclofen, and/or intramuscular botulinum toxin injections . As there is no specific treatment for the LND-associated SMB, these children often require removal of teeth as well as physical restraints including masks against spitting .
Since 2001, no more than nine LND patients treated with DBS have been reported. All patients were male, in line with the X-linked inheritance. The majority were operated in early adolescence (median 13 years, range 5–28 years). Exacerbation of SMB and severe dystonia were the principal indications for surgery.
In the literature, only the GPi has been used as a DBS target for LND. An interesting question is whether or not stimulation of the limbic GPi (with connections to the frontal cortex ) specifically addresses SMB, wheras stimulation of the motor GPi (with connections to the motor cortex ) improves dystonia. Two groups have implanted two electrodes on each side, one targeting the motor GPi and one targeting the limbic GPi [84, 85]. The motor and limbic GPi may be discriminated electrophysiologically , and Cif et al. indeed report specific improvement of SMB with limbic GPi stimulation (but not with motor GPi stimulation) and dystonia with motor GPi stimulation (but not with limbic GPi stimulation) . However, SMB can also be substantially or completely abolished when only one electrode is implanted, targeting the ventroposterior motor GPi . It seems that stimulation in one part of the GPi also affects neurons in the other part, as the firing rate of motor GPi neurons altered with stimulation of the limbic GPi . In LND patients, a firing rate of 10–15 Hz of has been reported in both motor and limbic GPi neurons, under propofol or sevoflurane anesthesia [76, 85, 86].
Programming and Follow-Up
In the published cases, only high-frequency stimulation (≥120 Hz) has been reported, although the pulse width varies widely (60–450 µs) between groups.
The most disabling symptoms in LND are SMB and dystonia, and both are improved by DBS. Dystonia severity (BFMDRS-M) is reduced by 4–55% (median 33%), 1 year postoperatively. The effect on SMB is impressive, with complete resolution or significant improvement reported in 6/8 and 2/8 patients, respectively. Both dystonia and SMB improve within the first days until 2–3 months postoperatively [5, 75–77, 84]. When stimulation is ceased, dystonia returns within 10 days, and SMB after 1–3 weeks . The evolution of BFMDRS-D 1 year post DBS was highly variable (ranging from 13% worsening to 50% improvement). However, many reports stress that physical restraints were no longer needed postoperatively [5, 75–77, 84] and that teeth extractions could be avoided . This is paramount for social interaction, going to school, oral communication, and feeding. Moreover, mood and level of cooperation can also improve . There have been no reports describing changes in cognition after DBS. Long-term experience is limited (with a follow-up of 2 and 5 years reported in three and one patients, respectively), but appears to be remarkably stable, both in terms of dystonia and SMB.
In 7/8 patients, dystonia was reduced substantially (>20%), while in 1/8 patients only a 16% BMFDRS-M improvement was obtained. Given the small cohort size and the heterogeneous way of outcome reporting, factors predicting responders cannot be reliably extracted. However, the improvement in SMB and dystonia is not larger per se in patients with four electrodes (targeting the motor and limbic GPi separately) than in patients with two electrodes. There is no obvious association with disease duration. Postoperatively, most reports describe a slow but substantial reduction in medication used to treat dystonia, and the complete cessation of anti-SMB drugs [75, 84, 85]. Given the selective effects of limbic and motor GPi stimulation, it seems plausible that DBS of the limbic and motor GPi alleviates SMB and dystonia, respectively. Nevertheless, SMB also improves with DBS targeting the motor GPi only. Interestingly, SMB may be a lateralized symptom, as it returned only on the left hemibody after a fracture of the right-sided GPi electrode . In 2/9 patients, a lead fracture occurred [15, 76], which is a relatively high rate and potentially associated with SMB or with stress on the leads due to growth. A late infection occurred in one patient, necessitating unilateral system removal . One sudden death, 1 year postoperatively, was reported .