Surgery for Parkinson’s Disease
Emad N. Eskandar, MD, G. Rees Cosgrove, MD, FRCS (C), and Leslie Shinobu MD, PhD

Mini Presentation: PD Surgery (ppt version) - or - PD Surgery (html version)

Address for Correspondence:
Emad N. Eskandar, M.D.
Massachusetts General Hospital
15 Parkman St. ACC # 331
Boston, MA 02114

E-mail: eeskandar@partners.org
Patient Appointments: 617.724.6590
FAX: 617.724.0339

Referrals | Stereotactic Surgery | Parkinson's Disease | Intractable Epilepsy | Movement Disorder Surgery
Guestbook | Selected Publications | Links

Introduction

The last decade has experienced a resurgence of interest in the neurosurgical treatment of Parkinson’s Disease due to simultaneous advances in clinical neurosurgery and basic neuroscience. The current model of basal ganglia connections and their role in movement disorders provides a rational basis for the neurosurgical treatment of PD. There are three targets for the neurosurgical treatment of Parkinson’s disease: the globus pallidus interna (Gpi), the subthalamic nucleus (STN), and the Vim nucleus of the thalamus. Options for treatment include the implantation of deep brain stimulators in one or more of these three areas (Gpi, STN, and VIM) or the creation of small lesion in Gpi (Pallidotomy) or the Vim nucleus of the thalamus (Thalamatomy). The choice of which treatment and the best target for treatment is based on a careful evaluation of each patient and their needs by our movement disorders team. Currently, the Gpi and STN are the preferred targets for the treatment of PD and the following discussion will be focused on these two sites. All operations are performed using intra-operative micro-electrode recordings to optimize electrode or lesion location for the best possible clinical outcome. What follows is a description of the scientific basis for the neurosurgical therapy of PD, the techniques we utilize at MGH, and a brief review of the indications and outcomes of the various techniques.

Basal Ganglia Anatomy

The basal ganglia normally comprise five nuclei (caudate, putamen, globus pallidus, substantia nigra, and subthalamic nucleus). The globus pallidus is further subdivided into the globus pallidus externa (Gpe) and globus pallidus interna (Gpi). The caudate and putamen together are called the striatum or neostriatum while the putamen and globus pallidus are called the lentiform nucleus.

The current model suggests that there are two pathways through the basal ganglia - Direct and Indirect. Both pathways begin in the cortex and both pathways converge on the globus pallidus interna, the main outflow of the basal ganglia. The direct pathway is thought to facilitate movements while the indirect pathway is thought to inhibit unwanted movements.

Direct Pathway

The direct pathway begins with projections from the cortex to the putamen. One population of putaminal neurons projects directly to the Gpi. The Gpi projects to the thalamus which then back to the cortex. The connections from the cortex to the putamen use glutamate and are excitatory. The connections from the striatum to the Gpi use GABA and are inhibitory, as are the connections from Gpi to the thalamus. The connections from the thalamus back to the cortex are excitatory. The cortex excites the striatum which then inhibits the Gpi. The Gpi is normally tonically active and inhibitory to the thalamus. When the Gpi is inhibited, the thalamus is relieved from inhibition and excites the cortex thereby reinforcing the desired movement.

Indirect Pathway

In the indirect pathway a separate group of striatal neurons projects to Gpe. Gpe then projects to the subthalamic nucleus. The subthalamic nucleus projects to Gpi. The Gpi projects to ventolateral thalamus and the thalamus projects back to the cortex. In contrast to the direct pathway the projections from the striatum to the Gpe use GABA and are inhibitory. The projections from the Gpe to the subthalamic nucleus also use GABA and are inhibitory. The projection from the subthalamic nucleus to the Gpi is excitatory. Therefore, the cortex excites the striatum which then inhibits Gpe. Since Gpe is normally inhibitory to the subthalamic nucleus, the subthalamic nucleus becomes more active and excites the Gpi. The Gpi being more active, then inhibits the thalamus and the thalamus does not excite the cortex. In this way, activation of the indirect pathway causes a relative inhibition of movement.

Parkinson’s Disease

This common disease was first described by James Parkinson in 1817. The disease occurs in about 1% of people over age 65. The peak onset is in the sixth decade of life. There are no proven genetic factors contributing to the disease. The cardinal manifestations of the disease are bradykinesia (abnormal slowness of movement), akinesia (absence of movement), rigidity (affecting extensors and flexors equally, often described as lead pipe or cogwheel rigidity), and tremor (fine tremor of the hands typically seen at rest, described as “pill rolling”). In addition, a number of other findings are associated with PD including masked facies (blank expression), festinating gait (patients walk with many small steps and have difficulty starting, stopping and turning), and micrographia (very small writing).

The primary derangement in Parkinson’ disease is a loss of dopaminergic neurons in the substantia nigra. The loss of dopamine results in derangement in both the direct and indirect pathways. However the effect of dopamine is different in the two pathways. Dopamine is thought to be excitatory to striatal neurons in the direct pathway thus it normally facilitates movements. Dopamine is inhibitory to the striatal neurons in the indirect pathway which also facilitates movements. The differential effect of dopamine on the two sets of neurons is due to the presence of different dopamine receptors in the striatal neurons. Striatal neurons in the direct pathway have D1 receptors which are excitatory. Striatal neurons in the indirect pathway have D2 receptors which are inhibitory.

In Parkinson’s disease, dopaminergic input is lost to both pathways. Therefore the direct pathway becomes less active and the indirect pathway becomes more active. In the indirect pathway, the loss of dopamine results in excessive activity of the subthalamic nucleus which leads to excessive activity in the Gpi. In both cases there is excessive inhibition of the thalamus by the Gpi which then presumably leads to the observed paucity of movements in Parkinson’s disease. The important point is that since both the Gpi and the subthalamic nucleus are overactive in Parkinson’s disease both of these nuclei are potential targets for surgical therapy when medical treatment has reached its limits.

Therapy of Parkinson’s Disease

The mainstay of therapy for PD has been a combination of l-dopa and carbidopa. Dopamine does not cross the blood brain barrier (BBB). L-dopa is a precursor of dopamine and does cross the blood brain barrier. Carbidopa inhibits the peripheral conversion of l-dopa to dopamine thereby allowing for more of the l-dopa to cross the BBB. Direct dopamine agonists can also be used in some cases. Dopamine is broken down by monoamine oxidase (MOA) and catechol-O-methyltransferase (COMT). Monoamine oxidase inhibitors can therefore be useful for treating PD. Amantadine is an antiviral agent which potentiates the effect of dopamine and is occasionally used.

After 5-10 years of treatment with l-dopa therapy patients often become less tolerant of the drug. Patients begin to cycle. During “OFF” periods, they have severe rigidity, akinesia, and tremor. During “ON” periods, they suffer from severe dyskinesias (involuntary movements of the limbs). For many patients medical treatment becomes increasingly difficult and it is nearly impossible to find a drug regimen that adequately controls the disease without side effects. Some of these patients are then candidates for neurosurgical therapy directed at either Gpi or STN.

CLINICAL AND TECHNICAL CONSIDERATIONS

Presurgical Evaluation of Surgery Candidates

Patients being considered for surgery should be evaluated by an experienced movement disorder team to ensure that they are good candidates for surgery and that all appropriate medical therapies have been tried. An essential element of this evaluation is to determine the major cause of the patients disability so that realistic goals and expectations can be agreed upon prior to surgery.

It is important to confirm the clinical diagnosis of idiopathic PD since a variety of neurodegenerative diseases can mimic PD in their early stages. Patients with these diseases appear to have a much poorer prognosis after surgery. Evidence of dementia or other cognitive decline, speech disorders, serious systemic disease, and advanced age are also considered contraindications to surgery.

Standard preoperative blood tests are performed with special attention to platelets, bleeding time, PT and PTT. Patients are advised to discontinue aspirin and other non-steroidal anti-inflammatory agents at least 5 days prior to surgery.

Anesthetic Considerations

Patients are kept NPO on the evening before surgery and are generally advised to withhold their anti-parkinsonian and anti-tremor medication on the morning of surgery. Surgery for Parkinson’s disease is performed under local anesthesia and requires the full cooperation of the patient therefore the intraoperative use of sedating agents is avoided. Intravenous access is established ipsilateral to the planned surgery to allow complete freedom of movement in the extremity of interest and oxygen is supplied by nasal cannula. EKG, pulse oximetry and BP is monitored but an arterial line is not inserted. Blood pressure should be maintained in the normal range for the patient. Bladder catheterization is not routinely performed.

Operative Technique – Pallidotomy or Deep Brain Stimulation

Frame Placement

A magnetic resonance (MRI) compatible stereotactic frame is affixed to the cranial vault after infiltration of the pin insertion sites with 1% lidocaine with 1/200,000 epinephrine. The pin insertion sites are chosen to avoid artifact through the planned axial imaging sections of interest. The frame should be placed as symmetrically as possible on the head to minimize rotation and lateral tilt. This ensures that any changes made in electrode position intraoperatively are entirely in the planned direction. We have utilized the Cosman, Roberts, Wells (CRW) frame (Radionics, Burlington, MA) but any MR compatible frame can be used.

Stereotactic Imaging

Following frame placement, the patient is taken to the MRI scanner where sagittal T1-weighted images are obtained first. These images are used to identify the anterior commissure (AC) and the posterior commissure (PC) and measure the AC-PC length. Next T1-weighted (TR 400, TE 12/Fr, FOV 30x30, 2 NEX, 3mm thickness) axial images are obtained through the basal ganglia so that these images are parallel to the AC-PC plane. Additional images in the coronal plane with fast spin echo inversion (FSE IR) recovery sequences to accentuate the gray-white matter borders of the thalamus and internal capsule can be utilized depending on the experience of the center. Alternatively, 3-dimensional volumetric acquisition series can be acquired which allow thinner sections and reconstruction in any plane but these series generally require 10 - 11 minutes and are thus subject to movement artifacts. While a variety of MR sequences can be used for targeting, it is imperative that each center verify the spatial accuracy of their stereotactic frame in their scanner with phantoms. Some inaccuracies can exist with stereotactic MR imaging due to field inhomogeneity and chemical shift artifact therefore we have also performed stereotactic CT imaging to enhance our targeting accuracy. Alternative techniques that employ image fusion software to combine stereotactic CT images with non-stereotactic MR images have also been used to successfully overcome these inaccuracies.

After MR imaging, the patient is taken to the CT scanner and after the appropriate localizer is placed, axial CT images (DFOV, 1.0 - 1.5 mm thickness) roughly parallel to the intercommissural line and through the area of interest are obtained. CT images, although geometrically accurate, do not provide the intrinsic anatomical details that MR images provide and should not be used alone for targeting purposes. In most cases however, the AC/PC line can be identified and used for initial target identification.

Target Selection

The initial target for pallidotomy or placement of a pallidal stimulator is 2-4 mm anterior to the mid AC-PC point, 19-22 mm lateral to the midline, and 4-6 mm below the AC-PC plane. The initial target for placement of a subthalamic stimulator is 3 mm posterior to the mid AC-PC point, 11-12 mm lateral to the midline, and 4 – 6 mm below the AC-PC plane. Although the spatial resolution of modern MRI and CT scanners continue to improve, detailed nuclear anatomy of the basal ganglia is still impossible to discern. Intraoperative physiologic confirmation of the target location through a combination microelectrode recording, stimulation, and the intraoperative assessment of clinical outcome remain essential.

Surgical Technique

Once the target point has been calculated, the patient is brought to the operating room. A single dose of an appropriate prophylactic antibiotic (typically cefazolin) is given. The patient is placed in a comfortable position and the frame is fixed to the operating table with the head only slightly elevated above the chest to avoid air embolism. A small patch of hair is shaved over the appropriate frontal region and the area is then prepared and draped. After infiltration of the scalp with 1% lidocaine, a 2 - 3 cm parasagittal incision is made and a burr hole placed 2.5 cm from the midline at the level of the coronal suture. The dura is coagulated and opened and a small pial incision is made avoiding any cortical veins to allow for atraumatic introduction of the electrode. At this point the stereotactic arc is brought into the target position and the electrode guide tube is lowered into the burr hole directly over the pial incision. The skin is then temporarily closed around the guide tube with nylon sutures to prevent excessive loss of CSF and brain settling.

Since there are individual differences in anatomy, initial target selection is only approximate and final targeting must be performed using micro-electrode recordings, macrostimulation or a combination of the two techniques combined with the clinical response of the patient. In general, the initial pass is performed with a micro-electrode to obtain spontaneous neurophysiologic data and then this is followed by the macrostimulation/lesioning electrode.

Microelectrode Recording

Microelectrode recordings are performed using fine high impedance microelectrodes (0.3 – 1.0 Mohm) which allow isolation and recording of extracellular action potentials. Both kinds of electrodes are then connected via short leads to a preamplifier. The signal from the preamplifier is then filtered, amplified, and passed through a window discriminator. There are several reasons for using microelectrode recordings as an aid to localization. Borders between white and gray matter are easily identified as white matter is usually very quite in comparison to gray matter. Different basal ganglia nuclei have different patterns of activity which can serve to reliably distinguish among the nuclei. Motor territories of the different nuclei can be determined because the isolated neurons are modulated by movement of the contralateral limbs. The spatial resolution of microelectrodes is excellent allowing for precise determination of nuclear boundaries.

Microelectrode Localization of Gpi

On a standard approach to the Gpi the microelectrode typically traverses the putamen, the globus pallidus externa, the globus pallidus interna, the ansa lenticularis, and finally the optic tract. The activity in these areas is characteristic and is useful in ensuring that the eventual lesion or stimulating electrode are in the correct location. Neurons in the putamen generally have a low baseline activity (0-10 hz) but can be modulated strongly by specific movements in the contralateral limbs. Neurons in Gpe have two distinct patterns of activity. Some units have a slow frequency discharge (10-20 Hz) punctuated by rapid bursts. Other units discharge with an irregular pattern at higher frequencies (30-60 Hz) but also have a bursting pattern with intervening periods of low activity. Many neurons in Gpe respond to repetitive movements with the majority of cells showing an increase in activity to passive or active movements of the contralateral limbs. In contrast, neurons in the Gpi of Parkinson’s patients have a higher baseline firing rate than neurons in Gpe (mean 80 Hz) . Furthermore, neurons in Gpi have few of the pauses in activity observed in Gpe. Most commonly, Gpi neurons respond to contralateral movements with an increase in firing rate. In addition, some neurons have been found that respond in synchrony with the patients’ tremor. The motor portion of Gpi is identified by the presence of neurons which are modulated by movements of the contralateral limbs. Once the electrode tip exits the floor of the Gpi there is usually a marked decrease in cellular acitivity. The optic tract can be identified by recording visually evoked potential or be noting changes in the background in response to a flash light. We usually rely on macrostimulation to identify the optic tract.

Microelectrode Localization of Subthalamic Nucleus

On approach to the STN the microelectrode encounters the anterior nucleus of the thalamus, the zona incerta, the subthalamic nuclues and finally the substantia nigra. The activity of these areas is distinctive and is invaluable in identifying the STN. Recordings in the anterior nuclues of the thalamus reveal neurons which appear to be sparsely distributed, have relatively low firing rates 10-30 Hz and are not clearly modulated by contralateral movements. As the electrode advances through the zona incerta there may be a zone of relative quite. Recordings in the STN are dramatically different revealing multiple high frequency neurons which are difficult to isolate and have a firing frequency of 20-50 hz (See Figure). If the electrode is in the motor portion of the STN the neurons will be modulated by movements of the contralateral limbs, usually the proximal joints such as the shoulder or hip. Once the electrode exits the STN there may be another 1-2 mm of relative quite. The electrode then traverses the substantia nigra where neurons with relatively high firing rates 50-70 hz are identified. However, in contrast to STN the cellular density is less and correspondingly there is less background noise and single neurons are easier to isolate.

Neurons recorded intraoperatively from the STN in a patient with Parkinson's Disease.

Macrostimulation

Macrostimulation can also be used to delineate the optimal target location. A commercially available lesion generator (Radionics, Burlington, MA) is used for impedance monitoring, stimulation and as necessary, lesioning. A macroelecrode with a 2 mm uninsulated tip is introduced through the guide tube under impedance monitoring. The impedance is seen to drop about 100? when the gray matter of the basal ganglia is reached. The electrode is stopped at a point 6 mm above the target and macrostimulation is then used to further delineate the optimal target location. Low frequency stimulation is performed with square wave pulses at a frequency of 2 Hz at 0-5 Volts to obtain motor thresholds in order to insure that the lesion does not injure the internal capsule. High frequency stimulation using square wave pulses of 75 Hz at 0-5 Volts is used to assess for proximity to the optic tract, speech dysfunction, and amelioration of symptoms. Stimulation is carried out at 6 mm, 4 mm, and 2 mm above the target and at the target. At each point both low and high frequency stimulation is performed. Typically the motor thresholds are around 4-5 volts at the highest electrode position and decrease to about 2-3 volts at the target.

High frequency stimulation usually causes an improvement in contralateral rigidity and bradykinesia which may be readily appreciated intra-operatively by using tasks such as finger tapping, rapid pronation/supination of the forearm, and toe tapping. On occasion, high frequency stimulation may elicit dyskinesias, a finding which generally predicts a successful outcome. Potential problems in speech are also assessed during high frequency stimulation by asking the patient to repeat several complex phrases and noting any difficulties. Once the electrode is 2 mm above target visual thresholds are obtained by turning off the room lights and asking the patient to report if he sees any flashing lights as the voltage as quickly increased and decreased with the high frequency stimulation. The classical response is a perception of flashing lights or phosphenes in the contralateral hemifield. The minimal voltage which elicits visual phenomenon constitutes the visual threshold. The electrode is then lowered to the target position and visual thresholds are again assessed. If the electrode is correctly placed, visual thresholds are usually between 2-3 volts. At final target coordinates, the optimal motor thresholds should be 2-3 Volts while the visual thresholds should be greater than 2 Volts to avoid injury to the internal capsule or optic tract respectively.

Stimulator Placement

Currently, deep brain stimulation is approved by the FDA only for use in the thalaumus. A stimulator can be used in either the Gpi or the STN but it is considered an "Off Label" indication. There two models in wide use – the Medtronic 3387 and 3389 quadripolar electrodes. The leads have four platinum iridium contacts 1.5 mm in length separated by 1.5 mm (3387) or 0.5 mm (3389). The electrodes are connected to a pulse generator which is placed infraclavicularly. The generator may be programmed to perform any combination of monopolar or bipolar stimulation. Once the target coordinates have been obtained and refined by microelectrode recordings, the stimulating electrode is introduced to the appropriate depth. By temporarily connecting the lead to an external stimulator the inhibitory effect on tremor and rigidity can be assessed as can the presence of side effects. If all is well, the probe is secured in place using the supplied cap or bone cement. The primary incision is closed and the patient is then placed under general anesthesia. The infraclavicular pocket for the stimulator is made and the leads are tunneled and connected. Some groups use fluoroscopic guidance to ensure that the electrode has not migrated during the procedure.

Lesion Generation

In pallidotomy a radiofrequency lesion is created in the globus pallidus interna. Once the target has been confirmed by the appropriate physiology a test lesion is made at 46 ^o C for 60 seconds. During this time, the patient is tested neurologically for contralateral motor dexterity and sensation along with verbal skills. If there is improvement in tremor and no neurologic problems then a permanent lesion is made at 75 degrees for 60 seconds. During the lesioning, the neurological status of the patient is continuously monitored and lesioning is halted if any impairment or change is noted. If complete abolition of the tremor has not been accomplished, then the lesion may be enlarged as guided by the intraoperative physiologic responses and recordings.

Postoperative Care

Postoperatively, patients are allowed to take their next scheduled medication. After a brief period of observation patients are usually returned directly to their room. An MRI scan is obtained within the first 24 hours to assess the lesion location and to exclude perioperative complications. The patients continue with their preoperative medications and outside of mild analgesics, no other medications are usually necessary. Given that hemorrhage is a rare but important cause of serious morbidity, good control of blood pressure in the perioperative period is essential. On the first postoperative day the patient is assessed by the movement disorders team and if all is well, the patient is discharged that day. Sutures are removed one week after surgery. A short course of rehabilitation therapy may be indicated for some patients to optimize functional recovery of the affected limb.

Conclusions

The neurosurgical treatment of Parkinson’s disease has evolved substantially over the past fifty years. Initial enthusiasm for stereotactic ablative surgical therapy was followed by its abandonment and an almost complete reliance on medical therapy. More recently, there has been a resurgence of interest in stereotactic surgery. Current surgical therapies for PD include STN stimulation, Gpi stimulation, and pallidotomy. All of these can lead to significant clinical improvements in selected patients although the precise indication for the different techniques are still not clearly defined. Given the rapid advances in functional neurosurgery, it is likely the treatment strategies will continue to evolve. The optimal management of patients with movement disorders require a combined approach with medical therapy providing the first line of treatment and surgery providing an option for selected patients who can no longer be adequately managed with medical therapy alone. Whether alternative surgical treatments such neural transplantation can ultimately provide a long lasting cure remains to be seen.

Referrals | Stereotactic Surgery | Parkinson's Disease Intractable Epilepsy | Movement Disorder Surgery Guestbook | Selected Publications | Links