Clinical aspects

Description

DBS has frequently been compared to brain lesioning (ie, pallidotomy, thalamotomy, and subthalamotomy). The main advantages of DBS over ablation include the reversibility of the procedure, the ability to program the stimulator, and the ability to perform bilateral procedures without unacceptable side effects.

DBS pretreatment evaluation and screening. Patients considering DBS should seek a referral to an expert center for complete interdisciplinary screening. The pretreatment evaluation consists of three steps: (1) confirmation of the diagnosis, as well as confirmation that adequate medication trials have been performed; (2) surgical risk-benefit stratification through the use of an interdisciplinary DBS screening team; and (3) discussion of possible risks and benefits with the patient and family (64; 01). In addition, it is essential to discuss the physical and psychosocial expectations with the patient and family (33).

Because a significant number of DBS candidates will have been diagnosed, treated, and possibly even had a deep brain stimulator implanted at an outside institution, reconfirmation of the diagnosis is critical (66). Atypical parkinsonian syndromes may closely mimic idiopathic Parkinson disease but may not respond well to surgery. This is one of the main reasons why timing of DBS surgery is generally recommended at least four years from onset of symptoms in the case of parkinsonism. A report published in 2006 by a committee comprising of neurologists, neurosurgeons, and neuropsychologists on a consensus on DBS for Parkinson disease indicated the importance of neuropsychological testing as a prerequisite for DBS in patients with Parkinson disease. Cognitive function may decline following STN DBS in patients with existing neurocognitive decline as can mood disorders such as major depression (47).

Patients with major neurocognitive decline are, therefore, not ideal candidates for DBS, although DBS can still be considered for palliative reasons after discussions with patient and family.

In general, symptoms of postural instability do not improve with DBS therapy and can, in fact, worsen. For the reasons cited above, it is important to counsel patients regarding these possibilities and to set realistic expectations about what DBS can and cannot accomplish.

If a diagnosis of Parkinson disease is established, trials of maximally tolerated dosages of carbidopa-levodopa, dopamine agonists, and possibly other drugs should be administered before surgery. Other formulations of carbidopa-levodopa, such as extended-release formulations, can be used in patients with dyskinesia and motor fluctuations. Other strategies for medication adjustments may include a trial of amantadine for patients with dyskinesias or anticholinergics such as trihexyphenidyl for patients with medication-refractory tremors. Many patients can delay the need for surgery with adjustments and optimization of medications. Alternative surgical techniques for managing refractory symptoms can be discussed, such as a carbidopa-levodopa intestinal gel. A Parkinson disease patient should always undergo on-off levodopa testing, as the features that respond to levodopa (except for tremor) tend to improve following deep brain stimulation.

In Parkinson disease, for instance, DBS improves medication-refractory tremors and severe dyskinesias, but is not expected to improve patients with Parkinson disease to better than their best on-medication state.

An evaluation of a dystonic patient should focus on the identification of underlying etiologies. Some secondary causes of dystonia are less responsive to stimulation, though tardive dystonia responds very well. Imaging should be performed and can help exclude secondary causes of dystonia. Patients may be tested for the DYT-1 gene if the dystonia is generalized and onset was documented as occurring before 26 years of age. DYT-1 dystonia tends to respond well to DBS, but other isolated generalized dystonias, such as DYT-6, DYT-11, or idiopathic, may also respond well. The examination before dystonia surgery should confirm whether there is reducibility of contracted postures. This examination may, in some cases, require anesthesia, although this is rarely done. Contractures that fail to “reduce” by physical manipulation will not usually respond well to DBS surgery, though in some cases pain may be improved. A typical dystonia patient selected for surgical intervention should have failed trials of levodopa, anticholinergics, baclofen, and in some cases, VMAT-2 inhibitors. Earlier treatment with DBS in idiopathic or genetic generalized dystonias is associated with improved long-term outcomes (04). Focal dystonias should, in most cases, first be addressed by botulinum toxin therapy. Furthermore, there is evidence that cervical dystonia, tardive dystonia, and several other indications may respond favorably to DBS. For example, a review and metaanalysis reveal up to 40% to 70% improvement on motor scales with bilateral pallidal stimulation (56).

Essential tremor DBS candidates should in general fail medication trials with first-line therapies such as propranolol (beta-blockers) and primidone. Second-line agents may be used. Even with maximal medication management, tremor control in patients with essential tremor is about 50% and is largely limited by side effects (107).

Other movement disorders such as Tourette syndrome have been addressed with DBS, but less is known regarding proper targets, approaches, and screening. DBS is not FDA-approved for Tourette syndrome. Revised recommendations for DBS in Tourette syndrome have been published, and there is an international database from the Tourette Association of America that tracks outcomes (77; 22). Candidates should have failed medications, including an alpha-adrenergic agonist (clonidine), multiple dopamine antagonists (antipsychotics, tetrabenazine), and possibly a benzodiazepine. Comprehensive behavioral intervention therapy or habit reversal therapy should be offered to patients. The international registry tracks approach, targets, and outcomes (22). Patients enrolled in the registry had an average 45% improvement in their Yale Global Tic Severity Scale at one year from DBS implantation (52). Martino and colleagues published recommended principles or “pillars” for patient selection for DBS (53).

Other applications that are not FDA-approved for DBS include Huntington disease, tardive dyskinesia, and other tremors such as multiple sclerosis–associated tremors (15).

Cerebellar DBS has been attempted for dyskinetic cerebral palsy (DCP) and has shown promising results. This is an attractive target as it is not involved in hypoxic ischemic damage (17). It has also shown benefit in acquired dystonia, thus pointing toward the role of cerebellum in movement disorders and the cerebellum as a potential target for DBS in dystonia (14).

All potential DBS patients should undergo a general medical evaluation for cardiac, pulmonary, and other comorbid conditions such as diabetes, hypertension, or obesity that may increase surgical risk.

The final selection of the DBS candidates should always be based on a multi- or interdisciplinary evaluation, which may include a movement disorders neurologist, a trained and expert neurosurgeon, and a neuropsychologist. In many cases, psychiatry, speech and swallowing, physical therapy, occupational therapy, and social work consultations are helpful. Each member of the screening team should meet, interview, and examine the patient and bring the evaluation results to a team meeting. The team meeting should provide an exhaustive discussion of the risks and benefits, the safest approaches, and the best targets. The results should be shared directly with patients and families.

DBS procedure. Once the patient has been selected, a high-resolution volumetric magnetic resonance imaging (MRI) study is preoperatively acquired. If the scan is done close to the surgery, it may be used for targeting. Otherwise, the scan will have to be repeated. Many different imaging techniques can be utilized for identifying a stereotactic target for DBS surgery. Although MRI has a superior resolution, it can be subject to movement artifacts. Computed tomography (CT) has minimal movement artifact but lower resolution and is usually reserved only for cases in which MRI is contraindicated.

Advances in MRI technology are improving the accuracy of targeting. Tractography, which allows identifying the different white matter pathways, is helpful in preoperative planning and predicting side effects (60). Fast T1 gray matter inversion recovery sequence (FGATIR) improves target visualization by nulling the white matter signal. This particular sequence has been modified to null both fluid and white matter signal and is known as fluid and white matter suppression (FLAWS) sequence (84).

Using susceptibility mapping, evaluating iron deposition patterns can help inform targeting of the subthalamic nucleus (29). A similar approach using high-resolution MRI and tractography has improved the targeting accuracy of the thalamic ventralis intermedius nucleus (55). Also, robot-assisted stereotactic implantation is gaining greater acceptance, especially as it decreases implantation errors (69).

Many institutions use software to fuse MRI and CT images. MRI-CT fusion can also help save time on an operative day (eg, not having to lie in the MRI scanner with a frame fixed for 40 to 60 minutes). In addition, dopaminergic medications are frequently discontinued in Parkinson disease patients at least 12 hours before DBS surgery because medications can influence microelectrode recording and the clinical examination during the actual procedure.

There are multiple surgical techniques. In one, a stereotactic targeting frame is placed on the patient’s head, usually under local anesthesia. This headframe must be aligned with landmarks on the skull and remain on the patient throughout the procedure. The head frame acts as a reference system for stereotactic targeting. Stereotactic targeting is an exercise in “virtual reality,” whereby virtual 3-dimensional space from the patient’s brain images can be translated into the real space of the patient’s actual brain. The exercise uses reference points from the imaging (with the head frame) and the coordinates on the head frame. After “virtual” targeting through the use of specialized software, stereotactic coordinates can be identified, and the brain is converted into a mathematical Cartesian coordinate system. Each pixel on the image is assigned a location. The coordinates are chosen using the “surgeon’s eye” and the software and can be used in the “real space.” The coordinates are dialed into the head frame. Alternatively, frameless stereotactic approaches rather than frame-based stereotaxy may also be used.

Additionally, DBS has now been performed by using MRI-guided procedures that do not utilize microelectrode recording. Finally, some techniques may allow DBS to be performed awake or asleep, although there is considerable debate and ongoing research into which approach may be best for an individual patient. For example, an asleep robot-assisted deep brain stimulation implantation with intraoperative CT verification provides accurate targeting with 0.85 mm radial error (92). The frameless, MRI-guided, and asleep techniques may offer a better patient experience with a lower risk of hemorrhage (110). Most experts agree that the best approach is one in which the team has the most experience and one where the implanting team regularly reviews lead locations and outcomes.

The most commonly employed DBS electrodes remain quadripolar. Medtronic, Abbott (previously St. Jude Medical), NeuroPace, and Boston Scientific all make DBS electrodes. The Medtronic DBS system has received FDA approval for essential tremor, Parkinson disease, dystonia, and obsessive-compulsive disorder (the latter two are approved under an FDA humanitarian exemption). Abbott (previously St. Jude Medical) received FDA approval for essential tremor and Parkinson disease. NeuroPace is approved for closed-loop approaches in epilepsy (and will not be discussed further as this article pertains to movement disorders), and Boston Scientific received FDA approval for Parkinson disease and essential tremor. Individual devices consist of an implantable lead with four or more electrode contacts on the ventral (deep) tip. Abbott and Boston Scientific devices offer two segmented middle contacts that allow current steering in horizontal and sometimes vertical planes. The devices have an impulse generator (pacemaker), connecting wires (usually placed in the chest or abdomen), and patient remote control.

Unilateral, bilateral, and staged (one procedure/lead per operative setting) procedures for lead implantation may be employed. Unilateral and staged procedures may potentially lessen adverse events, minimize postoperative confusion, and limit cognitive dysfunction. Whether implanted on the same day or 2 to 4 weeks later, the impulse generator is usually not activated until brain swelling has subsided. However, as the field is moving towards constant current devices, this paradigm may be shifting. Constant current is less affected by changes in the brain tissue impedance from postoperative edema (74).

The different DBS systems available on the market offer different advantages. Okun summarizes the significant advantages and disadvantages of the currently available DBS systems (62). The different devices vary in battery life, rechargeability, capacity for current steering, current source, and logistical support systems. Invariably, a multidisciplinary approach will help make this decision, especially as the technologies continue to evolve.

Depending on the surgical team’s expertise, neurophysiological techniques can be employed to facilitate more accurate mapping of a brain target. The techniques utilized have been referred to as microelectrode recording, microelectrode stimulation, and macroelectrode stimulation. There are tiny sensorimotor, limbic, and associative regions within the typical basal ganglia nuclei used for movement disorders surgery (subthalamic nucleus, thalamus, and globus pallidus internus). These regions may have distinct somatotopy, meaning, for example, that the face, arm, and leg can be predictably identified and mapped in the sensorimotor region. To obtain the best possible outcome and avoid neuropsychiatric side effects, surgical teams have been careful to place the final electrode in the sensorimotor region of these small subnuclei for movement disorders indications. Thus, microelectrode recording and microstimulation techniques can facilitate the generation of a 3-dimensional physiological map of a DBS target. Also, expert teams try to avoid some white matter tracts and surrounding nuclei to minimize the side effects of stimulation. As DBS has evolved, though, we have reinterpreted movement disorders as disorders of circuitry or networks (49). As a result, investigational targets focusing on specific white matter tracts rather than “gray matter” nuclei are being tested for different indications.

Microelectrode recording involves driving platinum and iridium or tungsten microelectrodes with an approximate tip diameter of a human hair (measured in microns) into various brain structures. The microelectrode can be used to identify individual cell activity that may be unique to specific basal ganglia structures or regions. The information gathered from microelectrode recording can be used to identify the boundaries and map regions to generate a more precise picture of the deep nuclei and their particular location.

Techniques for microelectrode recording vary between groups. Some prefer to use a single microelectrode recording pass, whereas others use multiple individual or simultaneous passes to generate enough data to construct a precise three-dimensional map. No standard microelectrode recording technique exists. As soon as a location for the lead is determined (a few groups do not use microelectrodes and proceed to macrostimulation following targeting), the DBS lead can be placed. Macrostimulation can be performed to check for thresholds (benefits and side effects) and is also helpful to confirm the placement of the lead and its proximity to the surrounding brain structures. Macrostimulation in the operating room also helps determine usability and programmability and may lead the surgeon to change the electrode’s position in the brain. The neurologist or a physiologist is often present throughout the procedure to aid in the electrophysiological monitoring and interpretation of the map, and to examine the patient, who is usually awake and can provide helpful feedback.

Controversy exists as to the role of microelectrode recording in DBS. Proponents of this technique point out that accuracy at a few millimeters can only be achieved by using microelectrode recording and that this accuracy is crucial in improving outcomes. For example, current imaging techniques cannot accurately delineate the boundaries of a target on 1-millimeter slices. Similarly, somatotopy and identification of the sensorimotor regions in the target nuclei can only be reliably achieved by using microelectrode recording. Reasons to oppose microelectrode recording include a slightly higher incidence of intracranial bleeding, more extended postoperative hospital stays related to the number of microelectrode passes, prolonged surgery time, higher cost, and the need for skilled expertise. Existing data on the issue of microelectrode recording have been mixed. A large, prospective head-to-head trial is needed to evaluate the role of microelectrode recording in improving outcomes from DBS. As imaging techniques improve and methodologies evolve (direct MRI targeting), there may be less need for microelectrode recording in the future (93). Indeed, a retrospective evaluation comparing microelectrode versus intraoperative MRI-guided approaches revealed comparable clinical outcomes and safety profiles (10). Also, MRI-guided techniques forgoing microelectrode evaluation are increasingly used (81). Advances in imaging technology using diffusor tensor imaging technologies allow more accurate identification of affected circuits and improved clinical outcomes (106).

There may be an improvement in symptomatology before electrode activation regardless of the target, referred to as the “honeymoon” or microlesioning/implantation effect. This effect is difficult to account for in studies and clinical practice because stimulation-induced effects and washout periods can confound results. Nevertheless, these effects have been observed both in the operative suite and in the clinic. For example, in St. Jude's (now Abbott) randomized Parkinson disease DBS study, 25% of the patients were not activated for three months. There were motor benefits (on-off fluctuations) and also side effects (verbal fluency deficits) in the group without stimulation, highlighting the importance of the microlesion effects (55; 65).

Postoperative programming. Postoperative programming is typically administered during multiple follow-up outpatient clinic visits. The average patient requires programming four to eight times in the first six months following surgery; however, following six months post-procedure, adjustments become less frequent and appointments may become more centered on managing medications and other comorbidities. In addition, patients can go home with multiple potential deep brain stimulator programming settings, which may decrease the need for some of the in-person programming adjustments.

Most DBS devices at this point use constant current stimulation as opposed to constant voltage, as it is not sensitive to current changes in brain tissue impedance over time and, hence, allows for more uniform efficacy and side effect profile in Parkinson disease. The fluctuation in impedance is most pronounced in the first six months postimplantation, most likely related to tissue remodeling (105).

The timing for the initial DBS programming visit varies from center to center but usually occurs a few weeks after implantation in order to ensure that the microlesioning effect and edema from the surgery have dissipated. If Parkinson disease was the indication for surgery, patients are asked to attend visits in the off-medication state to eliminate bias and contributions from Parkinson drugs. During the initial programming visit, a physician or clinician programmer will usually test each lead’s contacts for clinical effects as well as for side effects. The contact with the most optimal therapeutic window (the setting that provides the most significant benefit with the most negligible side effects) is selected for stimulation. At the end of the visit, the stimulators are set conservatively. Patients can follow up multiple times over the next few months as stimulation is tuned and medications are adjusted. The frequency of follow-up visits depends on the DBS center protocol and the patient response. For optimal results, appropriate adjustments (unipolar vs. bipolar and changes in voltage/current, pulse width, and frequency) must be made by an expert or experienced programmer. Patients should be counseled that once optimal settings are achieved, it is more important over the long-term to modulate medications and to prescribe appropriate therapies than to chase changes in deep brain stimulator settings. Many patients with Parkinson disease, for example, will seek a setting to address disease-related changes in walking, talking, or thinking, which are likely beyond the reach of the technology.

The overarching goal of DBS for movement disorders is to improve a patient’s quality of life. Parkinson disease, essential tremor, and sometimes dystonia are progressive disorders, so the need for stimulation in these patients is unlikely to abate. Patients typically either remain stable in their settings or require slight changes to settings (few significant changes in settings are necessary for most patients following six months of stimulation, but minor “tune-ups” are routine). The electrodes are typically not removed unless there is an infection or hardware malfunction (lead fracture), in which case, the electrode(s) could potentially be replaced. In essential tremor, it has been shown that both disease progression and tolerance may develop and lead to worsening of tremors over time (68). Some patients with dystonia have a characteristic dystonic tremor that is usually managed using the same target(s) as essential tremor, notably the ventral intermediate nucleus of the thalamus (Vim). Dystonic tremor control in patients with Vim DBS mirrors that of those with essential tremor though the benefit for activities of daily living wanes after two to three years from implantation (86) This phenomenon is different than in Parkinson disease, where the tremor seems to remain suppressed long-term (32).

In the cases of dystonia and Tourette syndrome cases, there remains limited literature on stimulated patients, though many with dystonia will maintain improvements for more than a decade, and those with Tourette syndrome at least six years (70). Anecdotally, children with dystonia who outgrow their leads will worsen until re-implanted. In cases where DBS has not been effective, stimulation is typically stopped, but the electrodes are not usually explanted because there is an increased risk of more surgery.

Currently, DBS relies on constant fixed stimulation (called open-loop DBS) that can only be changed manually by a programmer. Adaptive or responsive DBS has been evaluated more recently. Adaptive DBS tailors the stimulation automatically to a physiologic signal. This responsiveness portends at least similar efficacy, decreased side effects, and lower energy usage, thus, more extended battery longevity (09).

Furthermore, segmented lead designs allow for directional DBS. This technique steers the current towards the desired areas and away from others associated with side effects. However, it does not obviate the need for an accurate surgical technique and lead position. Advanced programming techniques, akin to those used in non-segmented leads, can be used, such as bipolar, double monopolar, or interleaving-like programs (98; 83).

Indications

DBS for Parkinson disease. In Parkinson disease, DBS is considered when patients develop intolerable wearing-off, motor fluctuations, or dyskinesias despite optimal medical management. Other patients for whom DBS may also be considered include those with medication-refractory parkinsonian tremor or dopaminergic medication-induced side effects such as nausea and vomiting or impulse-control disorders. Strict criteria are in place for patients to qualify for DBS therapy, and only a small percentage of the Parkinson disease population will be suitable candidates at any single point in time.

An accurate diagnosis of idiopathic Parkinson disease is necessary for the chance at a successful outcome from DBS surgery. Dopaminergic responsiveness is one of the best indicators of a good outcome. Optimal surgical candidates are Parkinson disease patients who score poorly on the motor subset of the Unified Parkinson Disease Rating Scale (UPDRS), whose performance improves significantly (total 30% to 50% improvement in their overall motor score) with levodopa therapy. However, many patients with a less than 30% dopaminergic response may improve with DBS therapy. This discrepancy highlights the importance of interdisciplinary screening.

The subthalamic nucleus and the GPi are the two surgical targets that help the three cardinal symptoms (tremor, bradykinesia, and rigidity) of Parkinson disease the most (also helps dyskinesia). The ventralis intermedius nucleus of the thalamus may also be considered for Parkinson disease, but thalamic stimulation helps mostly essential tremor. Thalamic stimulation for Parkinson disease patients should only be considered if there is disabling postural-action tremor (not rest tremor). Stimulation of the pedunculopontine nucleus has been investigated for gait and freezing in Parkinson disease, but results have been mixed and not an approved target. Subthalamic nucleus stimulation was compared to GPi stimulation for Parkinson disease in a head-to-head randomized, multicenter trial. Both stimulation sites had similar improvements in the UPDRS motor score at 36 months (100). Patients undergoing subthalamic nucleus stimulation were able to be titrated to lower doses of dopaminergic medications postoperatively compared to those undergoing GPi stimulation. However, they were also noted to have mild worsening in performance on verbal learning testing and a more rapid decline in Mattis Dementia Rating scale scores at 36 months. There was more medication reduction in the subthalamic nucleus patients. These results suggested that either target is appropriate for treatment of the motor symptoms of Parkinson disease. However, selecting a specific stimulation site for a given patient may vary depending on the relative importance of specific nonmotor and medication-associated outcomes.

There have been several studies of DBS for Parkinson disease. First, bilateral subthalamic DBS was shown by a German multicenter group to be more effective than medication management in improving quality of life in Parkinson disease (27). A multicenter study from the UK (PD-SURG) confirmed these initial results (103). The NIH COMPARE study demonstrated that unilateral STN and GPi DBS had similar motor outcomes (63), and later a VA multicenter study (34) and a Dutch multicenter study (59) also confirmed similar results from the use of bilateral stimulation of either target. A comparison of the U.S. Veterans Affairs and Dutch multicenter studies reveals that deep brain stimulation’s major determinant is the preoperative levodopa responsiveness. Individuals with greater preoperative levodopa responsiveness have better improvement in motor and quality of life scores (26). A review of the results of the EARLYSTIM study has failed to reveal a relationship between preoperative levodopa responsiveness and outcome from DBS but rather a relationship between quality of life preoperatively and postoperative outcomes (78).

Bilateral subthalamic stimulation with a constant current device was shown to be effective against the quality of life symptoms, and this study showed that activated devices outperformed devices that were implanted, but not activated (65). One study showed that constant-current bilateral subthalamic stimulation was well tolerated except for mild decrements in verbal fluency on the Stroop task (90). A German and French multicenter study showed that earlier implantation of deep brain stimulators, particularly in patients who develop Parkinson disease below the age of 60, may be beneficial for the treatment of motor fluctuations (79). A European study of the Boston scientific device applied bilaterally in the subthalamic nucleus target showed similar results to the other major studies (96).

DBS for essential tremor. Essential tremor is one of the most commonly encountered movement disorders in clinical practice. It is characterized by a postural-action tremor involving the hands and sometimes the head or voice. Unfortunately, only 50% of patients with essential tremor can be successfully managed with medication alone, and many medications have cognitive and mood related side effects. Therefore, surgical therapies such as ablation or stimulation of the contralateral nucleus ventralis intermedius have become mainstream to address disabling tremor.

Patients with essential tremor should not be considered for DBS unless their tremor is disabling and unresponsive to standard medical therapies. Before surgery, patients should be deemed medication refractory by a movement disorders expert and should be evaluated by a multidisciplinary team. DBS of the nucleus ventralis intermedius has been employed as a treatment for other forms of medically refractory tremor, including tremor secondary to multiple sclerosis and dystonia. Essential tremor has been addressed by some groups through the use of zona incerta or subthalamic stimulation; however, ventralis intermedius stimulation remains the most common approach in clinical practice. Patients should be counseled that tremor may progress despite DBS, though positive benefits have been reported at three or more years postoperatively (32; 68).

DBS for dystonia. DBS is considered in dystonic patients with significant disability or in those who have failed medical therapy. These patients should have a diagnosis of dystonia made by a movement disorders neurologist. Both primary and secondary dystonias can be considered for surgery, though secondary dystonias appear to have generally lower rates of response, except for tardive dystonia. Dystonia is classified as isolated when it occurs without structural brain abnormalities or other neurologic signs. Secondary dystonias occur as a result of CNS lesions, which may include those resulting from trauma, stroke, infectious causes, or neurodegenerative disorders. Tardive dystonia is a type of secondary dystonia that occurs after exposure to medications that block dopamine receptors. Medications that should be tried for dystonia before a surgical intervention include anticholinergics, benzodiazepines, muscle relaxants, and botulinum toxin therapy.

The posteroventral lateral globus pallidus internus is currently the target of choice for the treatment of isolated generalized dystonia, though subthalamic nucleus and other targets have been emerging particularly for isolated segmental and cervical dystonia. The stimulation parameters used are in general higher than those in Parkinson disease, and neurostimulator battery changes may, therefore, be more frequent, and the use of rechargeable devices more cost-effective. Reports have shown that low-frequency (60 hertz) stimulation may be effective and may result in less battery drain, though studies have revealed that this approach does not work for all patients. Reports of nucleus ventralis intermedius, ventralis oralis anterior/ventralis oralis posterior, centromedian nucleus, subthalamic nucleus, cerebellum, and other dystonia targets have also been published in the literature with mixed results. Most experts agree that isolated dystonias and tardive dystonias have the best response to treatment. Most experts also agree that intervention before the onset of contractions is an important tenet of treatment with a surgical approach (07). The appropriate age for surgery is still controversial, especially as many isolated dystonia patients have pediatric onset.

DBS for Tourette syndrome. Gilles de la Tourette syndrome is characterized by multiple motor tics and phonic (vocal) tics lasting longer than a year and occurring before the age of 18. Many patients also have behavioral issues, such as attention deficit hyperactivity disorder, obsessive-compulsive disorder, and self-injurious behaviors. Although DBS is not approved by the FDA for Tourette syndrome, there have been encouraging reports and randomized studies of tic improvement from DBS in severe, medication-refractory Tourette patients. Specific targets have included the midline thalamic nuclei, such as the centromedian-parafascicular complex, the anterior limb of the internal capsule/nucleus accumbens, the ventral and posteroventral globus pallidus, as well as the subthalamic nucleus (89). The Tourette Association of America International Deep Brain Stimulation registry and database sheds light on approaches and outcomes in Tourette syndrome DBS (77; 22). Martinez-Ramirez and colleagues reviewed the clinical outcomes at one year of 185 patients enrolled in the database and registry. They observed a statistically significant reduction in the motor subscore from 21 to 12.91 and in the phonic subscore from 16.82 to 9.63. Adverse events developed in 35.4%; most were mild. Notable side effects include intracranial hemorrhage (n=2), infections (n=4), and lead removal (n=1) (52). In another study from the database and registry, the imaging data were analyzed in 110 patients. The major takeaway from the study was the notable variability in in-target location as “[t]here were regions within and surrounding [the globus pallidus internus] GPi and [centromedian] CM thalamus that improved tics for some patients but were ineffective for others” (41).

Other indications. Other applications that are not FDA-approved for DBS include Huntington disease, tardive dyskinesia, and other tremors such as multiple sclerosis-associated tremors (15).

Cerebellar DBS has been attempted for dyskinetic cerebral palsy (DCP) and has shown promising results. This is an attractive target as it is often not involved in hypoxic ischemic damage (17).

It has also shown benefit in acquired dystonia, thus pointing toward the role of cerebellum in movement disorders and the cerebellum as a potential target for DBS in dystonia (14).

Levodopa resistant freezing of gait (FOG) is a frequently encountered problem in clinics and is significantly disabling, especially because it increases susceptibility to falls. Conventional DBS targets provide minimal, if any, help with this phenomena and in fact may exacerbate this in some cases (45).

The ability to maintain locomotion relies on a complex neural architecture comprising of the primary motor cortex, supplementary motor area, the basal ganglia, thalamus and the mesencephalic locomotor region (MLR) cerebellum, and the spinal network of central pattern generator (85).

The mesencephalic locomotor region in the brainstem is generally believed to play a vital role in locomotion and helps integrate the sensorimotor and emotional stimuli that accompany maintenance of posture and gait. It comprises of the cuneiform nucleus and the pedunculopontine nucleus.

There has been an interest in the stimulation of the pedunculopontine nucleus for freezing of gait in Parkinson disease. The pedunculopontine nucleus is cholinergic and glutamatergic in the pontine tegmentum, and is thought to potentially be involved in gait freezing and postural instability in Parkinson disease. An open-label study of four patients with both pedunculopontine nucleus and subthalamic stimulation showed similar improvement in UPDRS motor scores to subthalamic stimulation alone, but increased regional cerebral blood flow to the midbrain, thalamus, globus pallidus, and cerebellum. In a study involving 11 patients who received pedunculopontine DBS (with or without subthalamic DBS), nine patients were evaluated and had an improvement in the freezing of gait at a low stimulation frequency of 10 to 30 Hz. This stimulation benefit would wane after four to six weeks of stimulation, requiring a cyclic stimulation by turning the DBS off at night (35). In a systemic review and metaanalysis, pedunculopontine DBS improved axial symptoms though the results showed significant heterogeneity. One major factor for this heterogeneity was a notable effect of time since surgery, as there was an improvement in axial symptoms within the first year that was not noted beyond one year (48).

DBS of the cuneiform nucleus for levodopa resistant freezing of gait has been proposed but is yet to be studied (18). Cuneiform nucleus is located in the mesencephalic locomotor region (MLR), which is located in the midbrain tegmentum and has been found to initiate locomotion in cats.

Substantia nigra pars reticulata stimulation with microelectrode located in the STN and variable contacts in the SNr has shown benefit of gait in various studies, the hypothesis being that the SNr has connections with the mesencephalic locomotor region (102). Most of these studies also revealed that stimulating the STN and the SNr at different frequencies helped most with the gait freezing (91).

Contraindications

As previously mentioned, patients considering DBS surgery for their parkinsonian symptoms should be aware that a diagnosis of an atypical parkinsonian syndrome (eg, progressive supranuclear palsy, multiple system atrophy, dementia with Lewy bodies, corticobasal degeneration, or vascular parkinsonism) is usually a contraindication, though in select cases tremor, dystonia, and dyskinesia can be addressed. Atypical parkinsonian disorders are often referred for DBS because of a poor response to medical therapies but, unfortunately, typically have a less than optimal result from surgery.

Additionally, patients with Parkinson disease with primary goals to improve speech, respiratory and pharyngeal control, postural instability, and freezing may also be poor candidates for the current DBS technologies and targets. Several long-term studies have shown that dopaminergic-resistant symptoms continue to progress despite changes in stimulation parameters and improvement of other motor symptoms, such as tremor, rigidity, and bradykinesia.

DBS in Parkinson disease is usually performed to improve motor symptoms, though studies have revealed a smattering of improvement in some nonmotor features. DBS is not typically utilized in patients with more than mild cognitive dysfunction or with active psychiatric disease. The presence of severe dementia or significant cognitive impairment is usually considered a contraindication. This reluctance is due to reports of cognition in Parkinson disease worsening irreversibly after DBS surgery in patients with preexisting cognitive impairment. Also, dementia has been an exclusionary criterion in all clinical studies of DBS. Most centers use detailed neuropsychological testing to screen for dementia. However, it should be noted that occasionally patients with cognitive issues may be candidates for DBS, and this underscores the importance of an interdisciplinary approach and the construction of a risk-benefit profile.

There have been conflicting studies on whether age affects outcomes in DBS for Parkinson disease, resulting in most experts recommending DBS based on “physiological” age as well as most centers balancing comorbidities and cognitive function. DBS can usually be safely applied to cases above the age of 70 to 75 who have undergone careful screening and who understand the increased risks of surgery, mainly postoperative confusion (82).

Elevated suicide rates and depression post-DBS surgery for movement disorders have been reported. The reasons for this phenomenon remain unclear. Postoperative symptoms of depression and impulsivity may be due to serotonergic depletion following subthalamic nucleus DBS or mesolimbic dopaminergic withdrawal observed when patients are tapered off dopaminergic medications too rapidly (postoperatively). The risk factors leading to postoperative depression and suicide are unclear, but patients with preoperative depression or previous suicide attempts may be more likely to develop postoperative depression. Collectively, these findings have led to the recommendations of employing caution and close monitoring of patients with uncontrolled psychiatric illness or patients who may be at higher risk for suicide. Data from the VA multicenter study has suggested the possibility that with proper surveillance techniques, suicide may not be elevated post-DBS (101). Furthermore, patients with mild to moderate anxiety and depression did not have a higher risk of cognitive or psychiatric complications post-DBS (75).

Dystonia patients with fixed dystonic postures may not be candidates for DBS because fixed dystonic postures typically do not improve with surgery. Secondary dystonias, except for tardive dystonia, appear to have variable and generally lower rates of response. For DBS in Tourette syndrome, the patient’s motor and vocal tics should not be attributable to another medical, neurologic, or psychiatric disorder, and they should be the main drivers of impairment. We do not yet know if features beyond motor and phonic tics will improve consistently in Tourette syndrome DBS.

Although DBS implantation in pregnant women has not been studied, women of childbearing age can receive DBS, and most experts agree that DBS is not a contraindication to pregnancy. The obstetrician should be aware of the presence of DBS (to avoid diathermy), and it is usually recommended that women with DBS deliver at a hospital rather than at home (67). In a retrospective series of 11 patients with DBS (for multiple reasons), all pregnancies were uneventful except one complicated by a spontaneous early abortion of one fetus in a twin pregnancy. Five patients needed programming adjustments due to the worsening of their clinical symptoms. The authors have posited that the DBS surgery was essential for the success of those pregnancies (76). A similar report regarding six patients with dystonia treated with DBS did not reveal any concerns during pregnancy, delivery, or breastfeeding (109).

Medical comorbidities should be carefully addressed prior to consideration of DBS in any patient. Although not absolute contraindications, cardiac, pulmonary, and other comorbid conditions such as diabetes, heart disease, and obesity may increase surgical risk. Diabetes or other chronic medical conditions may impact wound healing and result in higher rates of infection. Those with uncontrolled hypertension have been shown to have an increased risk of hemorrhage with microelectrode recording, and the anesthesia and surgical teams must closely monitor this issue closely.

Outcomes

Parkinson disease. The commonly applied targets for DBS in Parkinson disease remain the subthalamic nucleus or the globus pallidus interna (GPi). In well-selected patients, they improve the three cardinal symptoms of Parkinson disease--tremor, rigidity, and bradykinesia--as well as motor fluctuations and dyskinesia. Unfortunately, dopaminergic-resistant features, such as speech, gait, and postural instability, are generally unresponsive to DBS when directed at the current targets. DBS has not been proven to be a neuroprotective therapy or to affect the ongoing loss of dopaminergic neurons positively.

DBS of the GPi or subthalamic nucleus can be powerful in treatment of motor fluctuations and dyskinesias. The effect lasts long term (studied more than 10 years in the subthalamic nucleus and more than five years in GPi) (39). Although the benefit of stimulation on motor symptoms is usually not better than the benefit obtained from dopaminergic medications, DBS can increase the amount of “on” time and reduce wearing off. In a randomized controlled trial of bilateral DBS versus best medical therapy, Parkinson disease patients who received stimulation in either the GPi or subthalamic nucleus gained 4.6 hours per day of “on” time without troubling dyskinesias whereas those patients managed with medication alone gained no “on” time (99). Dyskinesias have also been reduced with DBS in Parkinson disease. In a meta-analysis of subthalamic nucleus stimulation studies, dyskinesias were reduced by an average of 69.1% (44). Improvement of dyskinesias with subthalamic nucleus DBS is believed to be driven mainly by the reduction of dopaminergic medication after surgery, which can, on average, be reduced by about half. Pallidal stimulation has a potent antidyskinetic effect, and although medication reduction is less than subthalamic stimulation, there is more flexibility afforded by GPi DBS than subthalamic nucleus DBS. In the Deep Brain Stimulation for Parkinson Disease Study Group study, the 20 patients who underwent GPi DBS reported a 76% reduction in dyskinesias (23). However, most of these patients were not able to reduce the doses of their dopaminergic medications after surgery.

Subthalamic nucleus deep brain stimulation can improve freezing of gait at least in the short term (first year after implantation), and the effect appears mainly driven by improvement of other motor functions (such as bradykinesia and rigidity). Longer-term evaluation of the effect of subthalamic nucleus deep brain stimulation in the freezing of gait is limited. One study evaluated the freezing of gait in 52 patients who were evaluated up to seven years after implantation. Subthalamic nucleus deep brain stimulation appears only to improve the freezing of gait in the medication-off state and not in the medication-on state (43).

Quality-of-life measures have been shown to improve with DBS of the GPi or subthalamic nucleus for Parkinson disease, and some of these improvements can be sustained long-term, though many initial benefits in quality of life may be lost over time. These improvements seem to be driven by motor benefits. However, some evidence suggests that these changes may also be due to improvements in mood. The quality-of-life-dimensions that commonly improve after subthalamic nucleus stimulation include stigma, mobility, bodily discomfort, and emotional well-being, and the PDQ-39 scale used to evaluate post-DBS patients is heavily driven by mood changes. The neuropsychological outcome using constant current stimulation appears to be similar to that of the more standard constant voltage stimulation (90).

Below is a table that summarizes the outcomes of Parkinson disease DBS trials, adapted with permission from Dr. Michael S Okun:

Table 1. Outcomes of Parkinson Disease DBS Trials

Essential tremor. Studies have revealed improvements in tremor, disability, and quality of life after DBS for essential tremor. Unilateral thalamic stimulation reliably results in the marked improvement of contralateral action and postural and resting tremor and in some cases may even have a mild ipsilateral effect. Tremor control with DBS can be meaningful and long-term, but tremors may worsen over time in some patients, mainly due more to disease progression than to tolerance (32), though the control remains significant for most patients–estimated at 60.3% to 75% even with unilateral DBS (21).

Axial symptoms such as head and voice tremor are more difficult to improve with ventralis intermedius stimulation. Voice tremor is not reliably improved by unilateral ventralis intermedius stimulation. Bilateral ventralis intermedius DBS has been shown to reduce midline tremor compared to unilateral stimulation but may increase axial side effects such as dysarthria and gait instability. Indeed, for many patients, unilateral DBS may be enough for appropriate symptom control with 53% at 90 days and 63% at 180 days (54). One study showed that the trajectory of the DBS lead might impact the outcome of head tremors (57).

Vim DBS improves activities of daily living by up to 82%. Interestingly, most studies show that despite a more significant control of tremors with bilateral DBS compared to unilateral DBS, there is no significant difference in improvement in activities of daily living and quality of life (21).

As discussed previously, other targets can be used for tremor control, including the thalamic ventralis oralis nucleus, posterior subthalamic area, and the caudal zona incerta.

Dystonia. Primary generalized dystonias overall respond very well to DBS. This includes patients both with the DYT1 mutation (primary genetic dystonia) and those without the mutation (primary idiopathic dystonia). Kupsch and colleagues randomized 40 patients with primary dystonia to stimulation or sham stimulation for three months. At the three-month visit, patients who received stimulation had improvements of close to 40% in their Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS) subscores of movement and disability as compared to improvements of only 4.9% in BFMDRS movement and 11% in BFMDRS disability subscores from sham stimulation (46). The improvement increases to 57.8% at five years (97). Most case series of GPi DBS from primary generalized dystonia demonstrate improvements in BFMDRS scores between 50% and 70% with stimulation, and sustained motor improvement has been demonstrated at five years. Not all primary genetic dystonias respond similarly to DBS. For instance, TOR1A (DYT1), KMT2B, TAF1, and SGCE mutations have a better response to DBS in comparison to those with THAP1, ATP1A3, and GNAL mutations (08). Primary segmental dystonias may also respond to GPi stimulation. Most of these reported cases are cervical dystonia that is unresponsive to botulinum toxin therapy. Outcomes vary, ranging from 43% to 76% in clinical improvement using the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS). A metanalysis has pooled the results of two studies (total 102 patients) evaluating GPi DBS for patients with either generalized or cervical dystonia (73). With low quality evidence, this analysis lends support to the benefit of GPi DBS in controlling dystonia symptoms.

Despite these generally favorable group outcomes, not all patients improved with pallidal stimulation, even when lead placement was verified and thought to be optimal. It is unclear what factors predict a good outcome. Among patients with primary generalized dystonia, factors such as DYT1 status, short disease duration, and low severity of symptoms may predict a more favorable response. The presence of tonic posturing, fixed skeletal deformities, longer disease duration, greater age and lower GPi volume have been associated with worse outcomes.

Meige syndrome, characterized by blepharospasm and oromandibular dystonia, can respond to either GPi (about 60% improvement on BFMDRS and followed for more than ten years) or subthalamic nucleus DBS (74% improvement on BFDMRS and followed for up to four years) (40; 108).

Subthalamic stimulation has also been performed for primary cervical dystonia and generalized dystonia and was shown to have similar outcomes. For example, in cervical dystonia, a pooled meta-analysis that included 208 patients from 39 papers with either GPi or subthalamic DBS followed for an average of 23.3 months revealed a significant improvement in dystonia scales, severity, disability, and pain (87). Despite the lack of difference in clinical benefit, the side effect profile was different: parkinsonian bradykinesia observed in GPi DBS compared with dyskinesia and mood disorders in the subthalamic DBS cases.

Quality of life improvements have also been observed with GPi and subthalamic stimulation for primary dystonias. These improvements are most significant for physical quality of life measures, whereas cognitive measures were lower, possibly indicating the complex neuropsychiatric profile. Data on the quality of life of secondary dystonias is less reliable (88).

Secondary dystonias, except for tardive dystonia, appear to have variable and generally lower rates of response. A French multicenter group showed impressive improvements with GPi DBS for tardive dystonia. Birth, biochemical, developmental, medication, toxin, and trauma history are all potentially important points in these cases. The few larger case series on GPi DBS for tardive dystonia report clinical improvements in the 50% range. A French group led by Marie Vidailhet reported mildly positive outcomes in a group of patients with cerebral palsy undergoing DBS (95). Other targets, such as GPi (but not ventralis intermedius), improved symptoms in patients with dystonia-choreoathetosis cerebral palsy (104).

Table 2. Two Dystonia Studies that Demonstrate the Types of Improvements Observed in the Literature (Primary Dystonia)

Tourette syndrome. Early results have been promising, particularly for motor tics. In general, motor and phonic tic frequency can be reduced by about 50% in some published case reports and studies for both the thalamic (centromedian-parafascicular complex) and globus pallidus (motor and limbic) sites. However, it is unclear the extent to which the behavioral features will or will not respond. Stimulation of the internal capsule/nucleus accumbens may be more beneficial for the behavioral features of Tourette syndrome than for other sites though for tics results have been less impressive. Quality-of-life measures may also improve with thalamic DBS for Tourette syndrome. More information available from the Tourette Association of America International DBS Registry and Database has been published in the last couple of years, showing a significant improvement in motor and vocal tics that is persistent for at least five years. Interestingly, there is a delay in achieving clinical benefit with a median response time of 13 months (52; 41).

Adverse effects

There are three types of DBS-related risks or complications: (1) surgery-related, (2) device-related, and, (3) stimulation-related. Unfortunately, due to the lack of standardized guidelines for reporting adverse events, the published medical literature may currently underestimate the prevalence of DBS-related complications.

In addition to routine surgical risks, such as pulmonary embolism, pneumonia, bleeding from the wound or wound infection, postoperative seizures, and perioperative confusion, DBS presents about a 1% to 5% chance of intracerebral hemorrhage, but only 0.6% to 1.6% of implantations are associated with neurologic deficits (11; 42). The risk for hemorrhage seems to be higher with older patients and those with a history of hypertension, and increases with the number of microelectrode passes.

Hardware-related complications may occur in 9% or more of patients undergoing DBS, as shown in a retrospective review that included a mean follow-up of four years (06). Such complications include infection, device malfunction (electrode or wire break, IPG malfunction), skin erosion, and lead migration. The incidence of lead migration or lead fractures may be higher in the dystonia or essential tremor population, potentially because dystonic movements or a greater level of overall mobility can cause increased stress on the hardware. All patients with DBS devices should be warned to avoid diathermy treatment, as it can cause brain damage by induction of a radiofrequency current and heating of the electrodes. A patient developed a brain lesion due to an MRI of the lumbar spine causing the heating of the DBS electrode. However, MRI is believed to be safe in patients with implanted DBS devices if the manufacturer’s recommendations are followed. A review of the literature reported only four adverse events in the over 4000 published cases of MRIs in patients with implantable DBS hardware. Recent iterations of neurostimulators are no longer triggered by magnetic fields.

Stimulation-related side effects vary with the surgical target and with the surrounding neuroanatomy. These can usually be minimized by changing the stimulation parameters of the DBS system (re-programming). Stimulation-related side effects of ventralis intermedius DBS include dysarthria, paresthesias, dystonia, postural instability, ataxia, and limb weakness. These effects are noted more frequently with bilateral compared to unilateral stimulation and with left-sided rather than right-sided implants. Stimulation-related side effects with the electrode in the GPi region may include blurry vision or light flashes, dysarthria, paresthesias, or motor contractions. In contrast, stimulation-induced side effects in the subthalamic nucleus include dysarthria, paresthesias, motor contractions, diplopia, lightheadedness, sweating, dyskinesias, or hemiballismus. A subthalamic side effect of short- and long-term subthalamic DBS is brittle dyskinesia.

Indeed, a variety of movement disorders, such as blepharospasm and apraxia of eyelid opening, have been described with subthalamic nucleus stimulation, whereas hypokinesia and freezing of gait have been observed with stimulation of the GPi (05). Furthermore, ataxic gait has been described as a side effect of chronic bilateral stimulation of the ventralis intermedius. New and emerging technologies and stimulation techniques may help to prevent or overcome the various DBS-induced movement disorders. These include predictive models to guide surgeons in planning, steering/directional leads, novel pulse parameters (biphasic, variable frequency), and closed loop (or adaptive) DBS (72). Other side effects of subthalamic nucleus stimulation include apraxia of eyelid opening, weight gain, cognitive impairment, impaired recognition of emotions, and psychiatric symptoms such as impulse control disorders, depression, anxiety, apathy, and hypomania. Apraxia of eyelid opening is infrequent, and the mechanism is poorly understood. Many patients who develop apraxia of eyelid opening are treated with botulinum toxin. Weight gain with STN stimulation starts after surgery and may be significant. The underlying mechanism remains unclear but may be related to a reduction of energy output, improved alimentation, or direct influence on the function of the lateral hypothalamus.

Mood and cognitive effects of DBS have emerged as important, as they may limit the overall effectiveness of the therapy. Elevated suicide rates and depression post-subthalamic DBS have been reported. Though, with monitoring, the rates may be lower than originally thought (101). Indeed, unilateral subthalamic nucleus DBS appears to improve depression six months postoperatively (12). There is a possible effect of the location of stimulation on these mood changes. Hypomania has also been commonly seen, but does not tend to be sustained. Impulse control disorders may worsen after subthalamic stimulation, suggesting that the associative-limbic regions of the subthalamic nucleus may play a role in moderating the output of the striatal-frontal lobe circuitry. There have been many studies on the cognitive effects of subthalamic nucleus DBS; verbal fluency and executive dysfunction have been consistently reported to decline after surgery (63). The verbal fluency declines have been shown by multiple studies to be more surgically related than stimulation related (90). The high incidence of depressive and mood symptoms seen preoperatively in Parkinson disease makes the interpretation of postoperative effects tricky, but clinically relevant to track. There is a possible indication that GPi DBS offers a better long-term cognitive outcome than subthalamic nucleus deep brain stimulation, though more studies are needed (37).

The effect of DBS on the three cardinal Parkinson disease symptoms, tremor, rigidity, and bradykinesia, can be sustained for up to 10 years as long as the response to levodopa is also maintained. Quality-of-life improvements are also sustained long-term. Unfortunately, as mentioned earlier, dopaminergic-resistant features, such as speech, gait, and postural instability, do not respond to stimulation and continue to decline despite chronic long-term stimulation.

Several studies have been published confirming the efficacy of thalamic ventralis intermedius stimulation for essential tremor in long-term (1- to 6-year) follow-up. However, a small proportion of these patients may become refractory to DBS over time. This is discussed previously, and the debate persists on whether this is related to disease progression or tolerance/habituation to DBS.

There have been several case series examining long-term outcome of DBS for primary generalized dystonia and segmental dystonias. Side effects of stimulation are target specific. Globus pallidus internus is associated with bradykinesia and subthalamic DBS with dyskinesia, depression, and anxiety (87).

Despite the impressive motor outcomes of patients undergoing DBS for movement disorders, some patients have unsatisfactory results. In such cases of DBS “failures,” patients should have their lead locations confirmed by neuroimaging. If the lead locations are judged to be suboptimal, re-implantation has proved beneficial for many but not all patients. Stimulation parameter adjustment by an expert programmer at a DBS center may also improve results in up to 37% of patients. Unfortunately, 34% of DBS “failures” at an expert center could not be improved (66), with the most common reason for failure being misdiagnosis.

Because DBS does not arrest the progression of Parkinson disease, it does not seem to alter the long-term prognosis. Long-term data on the use of DBS for the treatment of essential tremor, dystonia, and Tourette syndrome are insufficient to support an alteration in disease prognosis.

Several clinical-pathological studies have confirmed that chronic DBS does not result in any significant pathology. In a comprehensive survey of 40 cases with 58 implanted DBS electrodes examined at autopsy, the mean age was 59.1 ± 13.0 (range: 21-88) years, and the mean postmortem histopathological follow-up was 22.2 ± 29.2 (range: 0.067-146) months (28). There were no significant pathological changes, but the examined brains showed fibrous sheaths (5 to 25 μm thickness) surrounding the electrode (94%), fibrillary gliosis (73%), reactive astrocytes (78%), multinucleated giant cells (75%), mononuclear leukocytes (92%), macrophages (91%), microglial activation (60%), axonal spheroids (64%), and mild neuronal loss (60%). To our knowledge, there are only four reports that discuss the occurrence of high-grade glioma in patients with chronically implanted DBS leads (02). This association appears to be coincidental and less likely causal given a large number of implanted leads and the minimal number of reported gliomas.

A significant advantage of DBS over ablative procedures such as focused ultrasound is its reversibility. Despite the possible development of chronic nonstimulation induced complications, DBS remains a largely reversible procedure (71).

Special considerations

Pediatric. DBS implantation in the pediatric population remains a controversial topic. Most of the current information regarding selection criteria, surgical targets, and programming are extrapolated from adult population experience. In dystonia, early implantation has been associated with better outcomes in most studies, though most of the evidence is based on low-quality data. However, there is concern for DBS lead migration due to brain growth. This migration can cause the retraction of the electrode by 5 to 10 mm between the ages of four and 18 years. Most brain growth occurs before the age of five years. It continues at a lower rate between the ages of five and seven years. Consequently, most experts recommend delaying DBS surgery in children until after seven years of age (50). Given the natural history of improvement in the majority of Tourette syndrome patients by adulthood, experts recommend a rigorous multidisciplinary evaluation, including local ethics committee involvement for DBS evaluation in pediatric patients with Tourette syndrome (77). A meta-analysis of pediatric DBS in Tourette syndrome revealed an improvement by 50% in 64% of the cases. Side effects were reported in 27.6% of children. Half of the side effects were surgical complications. Infectious complications occurred in 5.2% of the cases, which is higher than other pediatric DBS indications (20). Indeed, this trend of higher infectious complications has been reported in adult Tourette syndrome DBS and remains of unclear etiology (80).

Clinical vignette

This 64-year-old woman developed a right-hand rest tremor 17 years ago. The rest tremor was mild and intermittent, and she did not require treatment for five years. The tremor then became more consistent and was accompanied by loss of dexterity in the right hand, micrographia, and generalized slowness. She was initially started on selegiline, followed by low doses of carbidopa-levodopa, which offered her virtually complete control of her symptoms. She required higher doses of carbidopa-levodopa over the next few years as her symptoms spread to the left side of her body, with a prominent rest tremor of the left foot. Pramipexole and a long-acting carbidopa-levodopa preparations were added eight years before the presentation when she began to experience wearing-off symptoms. With the changes in medications, she noticed dyskinesias of her head and neck, which were mild until six years before her initial presentation. She retired as a school principal because of increasing problems with her Parkinson disease.

Over the few years prior to DBS evaluation, she had severe motor fluctuations. When the medications were working for her, she had uncontrollable dyskinesias. When she was “off” medications, she reported being almost unable to move. She estimated that she spent perhaps 10% of her day with uncontrollable dyskinesias, and more than 75% of her day with at least some level of dyskinetic movements. The rest of her waking day was spent in an "off" state.

She had no memory problems or hallucinations. Depression was always mild and well-controlled with paroxetine. Her speech was soft, but not dysarthric. She denied swallowing difficulty or sialorrhea. She had difficulty with activities of daily living because of her dyskinesias.

She had tried amantadine without a significant effect benefit for her dyskinesias. Entacapone did not increase her “on” time without increasing dyskinesias. Rasagiline was not an option because of potential interactions with paroxetine.

Past medical, surgical, and family histories were noncontributory.

On initial DBS evaluation, she scored 29 out of 30 on the Montreal Cognitive Assessment. Verbal fluency was 20 words in one minute. Attention, language, and higher cortical function were normal. Strength and sensation were normal. Reflexes were symmetric, and plantar responses were flexor bilaterally. She was examined in the “on” medication state with moderate generalized dyskinesias, no rigidity, and diminution in amplitude with finger-tapping, hand-grip, and pronation/supination movements.

She was brought back for motor evaluation after withholding dopaminergic medications since the previous evening. Her Unified Parkinson Disease Rating Scale (UPDRS) III score off medication was 30. After taking her usual dose of medication, her UPDRS III on medication improved to 10. She was appropriately counseled about realistic expectations from DBS surgery. After neuropsychological evaluation, speech evaluation, and evaluation by a DBS-trained neurosurgeon, the patient was presented and discussed at a multidisciplinary conference. It was felt that surgery was a reasonable option in this patient’s case, and she was scheduled for bilateral subthalamic nucleus implantation.

On the day of surgery, a stereotactic frame was placed under local anesthesia. A stereotactic MRI was then performed with the patient in the frame. Targeting was performed based on this MRI. The patient was then brought to the operating room. Microelectrode recording was performed bilaterally to localize the target region and to confirm optimal electrode tip placement. A DBS electrode was then placed after microelectrode recording and tested on both sides. Afterward, the DBS electrode was secured in the final location using fluoroscopic visualization. The pulse generator was implanted four weeks after the leads were placed. Initial programming was performed two weeks following the pulse generator implant. She required programming sessions every two weeks for the first two months and then every three months thereafter.

At her 6-month postoperative evaluation, she did not have any significant wearing off or dyskinesias. She was able to reduce her dopaminergic medications by greater than 50%. Her UPDRS III score in the off medication/on stimulation state was 12.

In summary, this patient had complications from her Parkinson disease medications, including dyskinesias and wearing off. She continued to have an excellent response to levodopa but did not have smooth motor function despite optimal medical management. She was appropriately counseled regarding risks and benefits and underwent a multidisciplinary screening and evaluation process before determining that she was an appropriate candidate. Most importantly, the patient was satisfied with her outcome at six months.

Scientific basis

Underlying anatomy and physiology. The anatomical targets of DBS vary with the disease and the symptom being addressed. However, as the earliest applications of DBS were to movement disorders, a discussion of the neuroanatomy involved in movement disorders is relevant. The basal ganglia serve to modulate many signals in the thalamocortical feedback loops that are involved in movement control. Aberrant neural activity in the basal ganglia and the afferent and efferent tracts connecting them to distant regions of the brain have been shown to be primarily responsible for the motor symptoms seen in movement disorders. Anatomical models of the basal ganglia have, thus, led to rational surgical treatments for these disorders. Ablative surgeries or the application of high-frequency stimulation with DBS to regions of the basal ganglia known to be involved in the control of movement, namely the thalamus, subthalamic nucleus, and globus pallidus interna, attempt to disrupt and to neuromodulate particular aspects of these thalamocortical loops and have successfully alleviated symptoms in many movement disorders, including Parkinson disease, essential tremor, and dystonia.

The "box model" represents the current dominant neuroanatomical theory of dopaminergic activity and circuitry connecting various nuclei in the basal ganglia. For many years, it was thought that there existed two major output systems of the striatum (the direct and indirect pathway) and that a balance between the activities of the two systems was necessary for neuromodulation, especially in the domain of movement. Several modifications and alternatives to this model have been proposed, including the addition of a “hyper direct” pathway between the subthalamic nucleus and cortex and a model in which Parkinson disease is the result of regionally aberrant firing patterns. A broader discussion of these issues can be found in this excellent review (25).

The “box model” successfully explains why damage to or stimulation of these basal ganglia targets improves parkinsonian symptoms. However, this model, although having some utility in explaining the pathophysiology of Parkinson disease and Huntington disease, is limited in that it fails to explain some other situations. For example, lesions in the thalamus or the external part of the globus pallidus do not produce clear akinesia or bradykinesia or disturbance of movement. Lesions to the external part of the globus pallidus also do not improve dyskinesias. Furthermore, destruction of the globus pallidus in the rat has a minor influence on the firing rate and pattern of subthalamic nucleus neurons (only a 19.5% increase). In contrast, depletion of nigral dopamine induced by 6-hydroxydopamine (6-OHDA) roughly doubles the subthalamic nucleus firing rate and also induces greater bursting and irregularity in neurons. Functional connectivity studies, such as functional MRI, have provided important insight into the altered network functions in Parkinson disease and other movement disorders. Unfortunately, there is much variability in the findings, and some evidence indicates poor reproducibility of the findings, which is most likely related to the heterogeneity of movement disorders (03). Fortunately, an expanding base of research in DBS may help to refine this model and to contribute to a better understanding of the circuitry connecting the basal ganglia in humans. For example, coherence of neuronal oscillation with the parkinsonian tremor frequency has been observed in Parkinson disease patients (38). Besides, Parkinson tremor causes a reduction of the beta band power from the subthalamic nucleus.

Given the new information, the “box model” has received multiple updates. Most notably, the excitatory or inhibitory functions of the direct and indirect pathways are no longer absolute. Both pathways have a mixed function with multiple bridging collaterals. An excellent review article summarizes these new updates and information about programming paradigms (30).

Mechanisms of action. The mechanism of action of DBS is not yet understood. What is known is that high-frequency DBS seems to in some unknown way override pathological burst activity, low frequency oscillations, synchronization, and anomalous firing patterns in nearby neural tissue. These signals are replaced with a still abnormal but less pathological signal that produces an often-dramatic improvement in motor symptoms. Although the exact driving frequency and mechanism likely varies between different diseases, there is a common theme of therapeutic effect through the overriding of pathological activity.

Clinical experience with DBS has revealed that minimal driving frequencies are required to produce therapeutic effects for a given disease and a given DBS target. The minimal frequency required to mask the intrinsic activity of a neuron in computer models increases as the frequency of intrinsic activity increases. Various nuclei likely possess varying intrinsic frequencies that can be overridden by DBS at different stimulation frequencies, some low and some high. It has also been noted that therapeutic range high-frequency stimulation and non-therapeutic range low-frequency stimulation of these targets can have markedly different clinical effects. Low-frequency stimulation in Parkinson disease in the 10 Hz range has either no clinical effect or even worsens Parkinson disease symptoms; improvement of motor symptoms begins to be seen in the 130 to 185 Hz stimulation range. In essential tremor, stimulation below 50 Hz has no affect or worsen tremors, whereas stimulation above 90 Hz improves symptoms (13).

Our knowledge of how high-frequency stimulation of DBS targets exerts its effects at the cellular level continues to increase. High-frequency stimulation inhibits the cell bodies of neurons that are close to the electrical field while directly and simultaneously exciting fibers, including afferent and efferent axons. In other words, it produces a functional lesion in a selected brain target region in a manner that mimics but is not the same as a lesion. The somatic inhibition may occur through activation of inhibitory afferent axons, presynaptic inhibition of excitatory axons, or depolarization-induced blockade of somatic ion channels. The DBS-induced excitation of efferent axons seems to be propagated to downstream nuclei (31).

More recent studies have demonstrated how these cellular effects may be contributing to the overall clinical changes resulting from DBS in Parkinson disease. It appears that the stimulation of afferent fibers to the subthalamic nucleus, likely from layer V of primary motor cortex M1, rather than stimulation of local neurons in the subthalamic nucleus, is at least in part responsible for the therapeutic effects of DBS in hemiparkinsonian rats. In these studies, selective inhibition of excitatory neuronal cell bodies in the subthalamic nucleus, direct activation of local glial cells, and directly generated oscillations in the subthalamic nucleus failed to alleviate parkinsonian motor symptoms. However, high-frequency stimulation targeted to afferent fibers to the subthalamic nucleus reduced spiking while allowing low-amplitude, high-frequency oscillations to persist in the local circuits and also reversibly and completely alleviate Parkinson disease symptoms. These same effects were seen with high-frequency stimulation targeting of layer V of primary motor cortex M1 projection neurons. As the subthalamic nucleus is a point of convergence for fibers from diverse regions of the brain, other afferents likely contribute to the therapeutic effects of DBS; however, it appears that at least the M1 layer V afferents participate in this response. Low-frequency stimulation of the afferent fibers increased beta-frequency firing in the subthalamic nucleus without affecting endogenous bursting and also worsened Parkinson disease symptoms. Unilateral subthalamic nucleus stimulation has also been shown to cause changes in the latency and frequency of firing rates seen in the contralateral subthalamic nucleus, suggesting that the benefits of DBS may be mediated through a more complicated cortical-thalamic network than is explained by a local functional lesioning effect.

An existing theory for why low frequency DBS fails to improve and at times exacerbates symptoms is that the long interpulse intervals fail to mask pathological activity in the stimulated neurons adequately and may even add additional spikes that further disrupt the activity of the region. The low-frequency stimulation burst may also stimulate rebound bursts in the stimulated nuclei or downstream nuclei, further promoting pathological activity. The effects of low-frequency stimulation should not be generalized to all nuclei and diseases. However, as low-frequency stimulation of the pedunculopontine nucleus may improve Parkinson disease symptoms as well as of the GPi for some subjects with dystonia.

Finally, not only the rate of neuronal firing but also the pattern of activity is important in the pathology of movement disorders. In the classic box model of the basal ganglia discussed above, the direct pathway initiates movements whereas the indirect pathway inhibits them. This model predicts increases in the rate of firing in the GPi in Parkinson disease, a hypokinetic disorder, and a decrease in the rate of firing in the GPi in Huntington disease, a hyperkinetic disorder. The rates of firing are, in fact, no different in these two diseases; however, the firing patterns and number of neurons recruited to participate in such patterns are different. It is agreed though, that this is in fact an oversimplified model and that the mechanisms by which movement is modulated are far more complex and involve different and concomitant cortical and subcortical networks.

In essence, DBS serves to modulate the neuronal circuitry via overriding of aberrant oscillatory activity. Previous studies have identified electrophysiological biomarkers for aberrant oscillatory circuitry such as the beta and the gamma band that constitute the local field potentials. There have been recent advances in technology directed towards identifying and recording local field potentials in real time so as to provide data for adaptive DBS in the future, also known as closed loop DBS. Medtronic has introduced the percept DBS system that is designed to serve this function and was recently FDA approved (36).

The ability to record local field potentials via the percept DBS would provide data that can be utilized for adaptive DBS, which appears to be a promising therapy for patients with motor fluctuations by delivering individualized stimulation targeting moment to moment changes in the neural circuitry (58).

A study by the Starr group found that when human subthalamic nucleus DBS was activated, there was a desynchronization of cortical activity as measured by phase amplitude coupling. This observation may provide a key piece to the puzzle when attempting to explain the mechanism of action of DBS (24).

Mahlon DeLong and Alim Benabid received the Lasker award for their work in this area and a summary of their accomplishments was published in the New England Journal of Medicine. We know a lot about the biological changes that result from applying an electrical current to the brain, but we still do not know which aspects of the biology underpin the mechanisms of action. Changes in neurophysiology, neurochemistry, neurovasculature, neurogenesis, and neuronal oscillations have all been documented post-DBS therapy.