Deep brain stimulation (DBS) is a procedure in which electrical wires are inserted into precise areas of the brain. These wires are then connected to a pacemaker-like device implanted subcutaneously in the chest. When the device is turned on, electrical stimulation is delivered to the brain and can improve the symptoms of certain conditions. DBS is an established therapy for Parkinson’s disease (PD) and essential tremor (ET), but has also been successfully used to treat patients with other movement disorders, such as dystonia and tics .
Tardive disorders are characterized by the presence of abnormal involuntary movements that occur after prolonged exposure to dopamine receptor blocking compounds, such as antipsychotic or antiemetic medications. The tardive disorders include tardive dyskinesia (typically a choreoathetoid or stereotypic movement involving the orobuccolingual region, but may involve other body parts), tardive dystonia (sustained contractions of agonist and antagonist muscles causing abnormal postures), tardive tremor (regular oscillation of a body part, such as the hand), tardive tics (sudden, jerky movements preceded by urges or sensations), tardive myoclonus (sudden muscle contractions producing a brief muscle twitch or movement), and tardive akathisia (subjective and observed restlessness) [2, 3].
Clinicians have been reluctant to use DBS for tardive disorders for at least three reasons. First, DBS for tardive disorders has not been tested in randomized controlled trials; only case reports and case series exist. Second, psychiatric side effects from DBS have been reported in other populations. PD patients treated with subthalamic nucleus (STN) DBS may develop depression, suicidality, mania, and impulse-control problems after electrode implantation [4, 5]. Thirdly, patients may develop psychotic delusions such as thought control, thought broadcasting, and thoughts of being controlled. Because the majority of patients who develop tardive disorders have underlying psychiatric disease (the exception is patients whose tardive syndrome is caused by a drug given for nonpsychiatric reasons), there is limited experience and evidence for DBS in tardive syndromes.
In recent years, several case reports and case series have demonstrated that DBS may improve tardive dyskinesia and dystonia in patients with and without psychiatric disease [6–30]. To date, DBS results have not been published for other tardive syndromes (akathisia, tics, tremors, myoclonus). Tardive dyskinesia affects 15% to 20% of patients treated with neuroleptics , while dystonia is present in 1% to 4% of patients exposed to dopamine receptor blockers . However, tardive dystonia often coexists with other tardive movements, so it is not uncommon to see patients with both tardive dystonia and the classic orobuccolingual movements of tardive dyskinesia [11, 14, 16], or patients with tardive dyskinesia and akathisia  or even blepharospasm . Because of this considerable overlap, and because the criteria for the diagnoses of tardive dystonia or dyskinesia are not always presented in these case reports, we will use the expression “tardive dyskinesia and dystonia” (TDD) in this chapter when the type of tardive syndrome is unclear or when reports are grouped together.
This chapter will discuss what is known about the mechanism of DBS, clinical outcomes from DBS, stimulation side effects, and surgical complications from DBS in patients with TDD.
The mechanism by which DBS is effective in controlling the involuntary movements of TDD remains unknown. In broad terms, there is abnormal signaling in the basal ganglia in patients with dystonia, and DBS somehow removes this abnormal signal, allowing better function of the basal ganglia . However, it still is not clear if DBS inhibits or activates the target nucleus. Some evidence supports activation of target nuclei [34, 35], while others show suppression or interruption of abnormal firing patterns .
Attempts to treat primary dystonia through neurosurgical interventions began in the 1950s. Initially, lesions were made, and though many different sites were tried, the GPi and especially the thalamus appeared to be the most frequent targets. Unfortunately, the effectiveness of these lesions varied markedly, and there were high rates of complication. The first report of DBS on dystonia was published by Mundinger in 1977 , when he used a thalamic target for cervical dystonia, but thalamotomies continued to be performed despite Mundinger’s good short-term results with stimulation. The success of pallidotomy for generalized dystonia in the 1990s switched attention back to the GPi, and several groups started successfully treating dystonia with GPi-DBS [38–40]. The efficacy and safety of GPi-DBS in primary dystonia is now well established , and in 2003, the US Food and Drug Administration (FDA) approved DBS for dystonia under a humanitarian device exemption. Because of this, and the decreased likelihood of behavioral side effects, the GPi has become the most common stimulation site for the treatment of TDD , though some cases of DBS in the thalamus and subthalamic nucleus have also been reported [28–30].
Prior to surgery, high-resolution volumetric magnetic resonance imaging (MRI) is usually performed. In many surgical centers, a stereotactic targeting frame is then placed on the patient’s head, and specialized software is used to identify stereotactic coordinates [43, 44]. Other centers may use a frameless stereotactic approach . Coordinates of the GPi target are typically around 20–22 mm lateral to midline, 2–3 mm anterior to the anterior commissure, and 4–6 mm below the bicommissural line . The DBS lead, impulse generator, and connecting wires are then implanted. Depending on the center, bilateral electrodes may be implanted on the same day, or the surgery may be performed in a staged fashion [46, 47]. The impulse generator and connecting wires may be implanted on the same day or later in a separate surgery.
Techniques such as microelectrode recording and macroelectrode stimulation may be used to enhance the accuracy of placement. Microelectrode recording identifies individual cell activity that may be unique to specific basal ganglia structures. This information can be used to map regions in order to generate a picture of the deep nuclei and their location. Techniques for microelectrode recording vary between groups. Some prefer to use a single microelectrode recording pass, whereas others use multiple simultaneous passes. As soon as a location for the lead is determined, the deep brain stimulation lead can be placed. Macrostimulation can then be performed to check for side effects. For example, with GPi placement, the lead might be too ventral if visual phosphenes occur with stimulation. If tonic motor contractions occur with GPi stimulation, the lead might be too posterior. The presence of side effects may lead the surgeon to move the lead . Electrode placement can be confirmed with postoperative imaging (MRI or CT). The timing for initiating stimulation varies from center to center, and may range from the day after to a few weeks after implantation.
Again, DBS has only been used to treat TDD, and the majority of cases have used the GPi as the target. The main criteria for undergoing DBS in these reports were 1) that TDD had to be disabling and refractory to medical management and 2) that the patient had to be stable psychiatrically and cognitively. The interval between exposure to neuroleptics and development of TDD varied from a few months to many years in the DBS cases, but in most cases, TDD had been present for many years before surgery .
Trottenberg et al. reported the first case of a tardive disorder being treated with DBS. This was a 70-year-old woman with a 6-year history of tardive dystonia, who had electrodes implanted in bilateral ventral intermediate thalamic nuclei (VIM) as well as bilateral GPi . With bilateral pallidal stimulation, the patient’s dystonia improved within hours. Bilateral thalamic stimulation did not improve symptoms, and the combination of thalamic and pallidal stimulation was not better than pallidal stimulation alone. Since then, a total of 66 patients have been reported in the literature who have undergone GPi-DBS for TDD . In general, patients have shown fairly consistent improvement in signs and symptoms based on standardized scales. It should be noted that all of these case reports were unblinded, with the exception of one report by the French Stimulation for Tardive Dyskinesia Study Group (STARDYS) . In this study, ten patients with severe tardive dyskinesia underwent bilateral GPi-DBS surgery. At six months, patients were evaluated in the stimulation on and stimulation off conditions in a double blind fashion (i.e., both patients and evaluators were blinded as to whether DBS was on or off). With stimulation on, patients experienced a mean improvement of 50% (range, 30%–66%) in symptoms based on the Extrapyramidal Symptoms Rating Scale.
Some description of the scales used in these studies is provided to better understand the outcomes. Three main scales have been used: the Burke-Fahn-Marsden Dystonia Rating Scale (BFMDRS) , the Abnormal Involuntary Movement Scale (AIMS) , and the Extrapyramidal Symptoms Rating Scale (ESRS) .
The BFMDRS is composed of two subscales: a motor subscale (based on an examination) and a disability subscale (based on patient report) . The motor subscale has nine components and assesses dystonia symptoms in the eyes, mouth, speech/swallowing, neck, arms, trunk, and legs. Both severity of dystonic symptoms and provoking factors for the dystonia in each body region are rated on a scale from 0 to 4. The score for each body region is obtained by multiplying the ratings of the severity and provoking factors, except for the eyes, mouth, and neck, where the product of the severity and provoking factors ratings is halved. Each body region score is then totaled, for a maximum score of 120. The disability subscale has seven items that assess speech, writing, feeding, eating, hygiene, dressing, and walking. All items are rated on a 5-point scale except walking, which is rated on a 7-point scale. The maximum disability subscore is 30. In general, the 12 patients in the literature who were evaluated with the BFMDRS total (motor plus disability) score showed a mean improvement that ranged from 6% to 100% and averaged 62.1% . A total of 42 patients also had the individual motor and disability subscores of the BFMDRS reported. These patients had a mean improvement of 71% in motor and 65% in disability subscore with GPi-DBS . The BFMDRS motor subscore can improve substantially, up to 87%, at 6 months after GPi-DBS , but there also have been cases of no response .
The AIMS is composed of seven items that rate the severity of involuntary movements from 0 (none) to 4 (severe) in the face, lips, jaw, tongue, extremities, and trunk . Tremor is excluded from the ratings. Items 8–10 refer to the patient’s global judgment of severity of symptoms, incapacitation due to abnormal movements and awareness of the abnormal movements. There are also two “yes/no” questions about dental status. The maximum possible score is 42. Spindler et al.  reported a man with severe tardive dyskinesia who experienced 71% improvement in the AIMS score after GPi-DBS, sustained for over 5 years with small adjustments to his stimulator voltage and medications. This level of response in the AIMS score with GPi-DBS is similar to that reported by a systematic review showing a mean improvement in the AIMS score by 71.5% (95% CI, 62.6%–79.3% [p<0.0001]) with GPi-DBS . In the double blind assessment by Damier et al. , the mean AIMS score improved by 56% (33%–69%) 6 months after GPI-DBS.
The ESRS is a scale that was specifically developed to rate drug-induced movement disorders . The questionnaire includes four subscales: 7 items assessing the patient’s subjective impression of their drug-induced movement disorder (I), 17 items based on examination of parkinsonism and akathisia (II), 10 items based on examination of dystonia (III), and 10 items based on examination of dyskinesia (IV). There are also four clinical global-impression-of-severity questions of parkinsonism, akathisia, dystonia, and dyskinesia (V–VIII) for a total of 45 items. In a systematic review, TDD improved by a mean of 67.2% (95% CI, 55.5%–77.2%) when assessed with the ESRS after GPi-DBS . A double blind evaluation of ten patients reported a similar (50%) improvement in the ESRS at the 6-month evaluation after GPi-DBS .
Though TDD overall seems to be responsive to GPi-DBS, different symptoms may respond differently. Fixed dystonic postures appear to be the least responsive to GPi-DBS [11, 16, 18, 20]. Phasic dystonic movements and orobuccolingual dyskinesias are quite responsive and tend to improve similarly [11, 18, 20]. Dystonic symptoms in different regions may also respond differently to DBS. Franzini et al. reported two cases where the tardive cervical dystonia appeared to have more improvement than orobuccolingual dystonia , while other investigators have reported that the degree of motor improvement appears to be equal in all anatomic regions (orobuccolingual, axial, and limb dystonia) [17, 19]. The response of axial vs. appendicular dyskinesias to GPi-DBS appears to be quite variable [7, 8, 15].
The time from stimulation initiation to benefit also varies. Some TDD cases have reported improvement within a few seconds after starting stimulation, while others may take months. The immediate response to stimulation was achieved in two cases of tardive dyskinesia who were unblinded as to whether the stimulation was on or off [6, 8]. For other cases of tardive dyskinesia, 10 patients improved within days, but the rest improved gradually over 3 to 6 months [7, 9, 15]. For those patients with tardive dystonia, phasic dystonic movements seem to respond more promptly to GPi-DBS, with improvement within days of stimulation, while tonic abnormal postures can take months to achieve significant improvement .
Long-term follow-up in patients undergoing DBS for tardive disorders is limited. From the 49 cases reviewed here, 29 patients were followed for more than 6 months and experienced sustained benefit as follow-up continued [7, 12–18, 42], in some cases for more than 6 years [13, 17]. See Table 17.1 for more details regarding follow-up.