Stereotactic ventral pallidotomy for Parkinson's disease

Dopaminergic therapy for Parkinson's disease (PD) is usually effective for 5 to 10 years but is often followed by deterioration with progressive bradykinesia, rigidity, tremor, and impairment of gait and balance ^[1]. Late-course deterioration is frequently associated with increasingly severe and often incapacitating levodopa-induced dyskinesias and fluctuations in clinical response ^[2,3].

Surgical attempts to control PD have a long history. Cortical resection, extirpation of the head of the caudate nucleus, and lesions of the anterior limb of the internal capsule have had limited success ^[4-8]. Improvement of parkinsonism with anterior choroidal artery ligation led to interest in lesioning the globus pallidus as a potential therapy ^[9]. The efficacy of pallidal procedures ^[10-13] was inconsistent due to problems with stereotactic localization ^[14,15] and the variable effects of chemotherapeutic agents ^[16]. Anecdotal cases and limited series reported that stereotactic pallidotomy improved parkinsonian symptoms ^[17-24] but they lacked objective outcome criteria or long-term follow-up. One series ^[25] reported that stereotactic posteroventral pallidotomy improves bradykinesia and balance in PD patients. In addition, primate studies ^[26] demonstrated that subthalamic nucleus (STN) lesions reduce bradykinesia in experimentally induced parkinsonism. Dopaminergic activity in the brain can be augmented alternaty by adrenal medullary and fetal tissue transplantation, ^[27-33] but transplantation techniques are complex and limited to a few centers, and ethical issues are unresolved. Against this background, we used state-of-the-art computer image-directed stereotactic destruction ^[34-37] of the ventral medial globus pallidus to improve the symptoms of medically intractable PD patients. This report describes the effects of stereotactic ventral pallidotomy on 18 patients.

Methods. Protocol and patient selection. Protocol approvals were obtained from the Institutional Review Boards at the Hospital for Joint Diseases-New York University Medical Center and the North Shore University Hospital-Cornell University Medical College. Specific informed consent was obtained for positron emission tomography (PET) and intraoperative neurophysiologic recording and stimulation.

A surgical group and a nonsurgically treated group were studied. The nonsurgical patients desired surgery but wished to wait until long-term results were available. Selection of patients was made by an advisory board of two neurologists and one neurosurgeon (E.F., E.K., and M.D.). PD was diagnosed when a patient demonstrated parkinsonism with response to levodopa and had no history of known causative factors such as encephalitis or neuroleptic treatment.

The following exclusion criteria were employed: (1) dementia, supranuclear gaze palsy, or ataxia; (2) stage IV Hoehn and Yahr ^[38] when "on"; (3) blood pressure drop greater than 30 mm Hg on standing; (4) medical, neurologic, or orthopedic illness that would compromise assessment, such as spinal stenosis, cerebrovascular disease, or metabolic disorder; or (5) the presence of a focal brain abnormality on MRI.

Surgical group. Eighteen patients with PD responsive to levodopa were studied. Inclusion criteria were (1) bradykinesia and rigidity-predominant PD, and (2) response fluctuations with "off" parkinsonism and "on" chorea/kinetic dystonia. The mean age was 59.8 years (range, 42 to 79). There were 11 men and seven women; mean disease duration was 10 years (range, 4 to 25).

Nonsurgical group. Seven patients with PD who met the same criteria as the surgical patients were studied. They were evaluated and followed in the same manner as the surgical patients. The mean age of patients in the nonsurgical group was 64 years (range, 43 to 77). There were six men and one woman; mean disease duration was 12 years (range, 7 to 17).

Patient assessment. Neurologic. Patients (18 surgical and seven nonsurgical) were assessed according to the "Core Assessment Program for Intracerebral Transplantation" (CAPIT) section III: definitions of OFF and ON, and section IV: core methodology A, B, C, and E (not D), ^[39] and by the Unified Parkinson's Disease Rating Scale (UPDRS) ^[40]. The CAPIT stand-walk-sit subtest was modified so that patients were timed walking 23 feet, turning, and then walking back to the start line.

The CAPIT assessment (four timed tests: pronation-supination, hand-arm movement between two points, finger dexterity, and walk) and the UPDRS (two subtests: Motor and Activities of Daily Living (ADL)) were administered after the patients had been off medications for 12 hours (12 HOM) and several hours after waking up, at the following time points: (1) at baseline, (2) within 1 week of surgery, and (3) every 3 months for the duration of the study. These tests were also administered to surgical patients in the "on" state (best "on" after administration of the first dose of medications following 12 HOM) at baseline and 1 year postoperatively. All patients were recorded on videotape, demonstrating their UPDRS Motor test and CAPIT performances. These recordings were made at baseline and at 3, 6, 9, and 12 months postsurgery (or postinitiation of study for nonsurgical patients). Surgical patients were recorded both at 12 HOM and in the "on" state, whereas nonsurgical patients were recorded only following 12 HOM. Baseline and 1-year follow-up tape recordings 12 HOM were randomized and blindly rated by two of the investigators (A.B. and D.E.), who were not involved in initial clinical assessment and follow-up of the patients.

Patients were maintained on a stable dose of antiparkinson medications for at least 2 weeks before surgery. This dose was maintained, when possible, after surgery.

Visual fields. All surgical patients underwent postoperative confrontational visual field analysis with a Goldmann computer-driven perimeter within 3 months after surgery.

Radiologic investigations. Magnetic resonance imaging. All patients (surgical and nonsurgical) had T₁- and T₂-weighted MRI. Surgical patients had stereotactic pre- and postoperative MRI for confirmation of lesion placement and size.

Surgical procedure. All patients were deprived of medications for 12 or more hours before surgery. Under local anesthesia, a Leksell stereotactic G-frame was affixed at four points using fiberglass pins, and the frame was applied to position the midpoint of the anterior commissural-posterior commissural (AC-PC) line in the center of the stereotactic frame (x = 100, y = 100, z = 100). The frame was centered in the MRI or CT scanner. The first 10 patients had both CT and MRI; thereafter, only MRI was used. Utilizing stereotactic MRI, a midsagittal section was obtained, and an AC-PC line and plane were established. Axial slices were obtained at 3-mm intervals to 12 mm below this plane. The data were digitized and transferred to an independent work station (CASS/Midco, San Diego, CA), which allowed correlations of fiducials in space and application of the appropriate Schaltenbrand and Wahren brain maps ^[41] to determine precise coordinates.

The anatomic target lay between 17 and 25 mm lateral to the midline, 6 to 8 mm below the AC-PC line, and 2 to 3 mm anterior to the midcommissural point (figure 1). A cranial opening was made with a 3.5-mm skin punch and 3-mm twist drill in the same plane as the x-coordinate (lateral to the midline) and at an angle between 28 degrees and 50 degrees to the AC-PC plane. Single-cell recording was performed with tungsten-tip, disposable microelectrodes (1 to 2 MOmega impedance at 1,000 Hz). We analyzed "on-line" firing pattern and single-cell responses to passive and active movement to physiologically confirm that the MRI-defined anatomic target was within the globus pallidus internus (GPi). The methods and results obtained during single-cell recording have been described in detail ^[42].

Electrical stimulation was performed before lesioning to prevent injury to the internal capsule or optic tract; the effects and stimulation thresholds eliciting contralateral muscle contractions or visual symptoms were recorded. Stimulation was performed utilizing a radio frequency lesioning electrode (Radionics) with a 1.1-mm diameter and an exposed 3-mm tip at pulse duration of 0.2 msec, 5- and 50-Hz trains, and intensities of stimulation of up to 10 V.

Multiple lesions were made at 80 degrees C for 60 seconds. Lesions were overlapped and created in sequence by withdrawing the electrode at 2-mm intervals, resulting in a cylindrically shaped lesion (figure 2).

Statistical analysis. The performance of surgical subjects was compared from baseline to postoperative follow-up (within-group analysis) and to age-matched, disease-matched, and antiparkinsonian medication-matched nonsurgical patients (between-group). ANOVAs and paired t tests were used to analyze the CAPIT scores. Three separate MANOVAs were performed for each of the postsurgical assessment periods. Each MANOVA contained the three timed scores derived from the CAPIT (contralateral upper limb, ipsilateral upper limb, and walk) as the dependent measures. Scores from the UPDRS were analyzed with nonparametric methods (Mann-Whitney U tests for between-group comparisons and Wilcoxon signed rank tests for within-group changes from baseline). The two blinded CAPIT results were compared with each other by paired t test, and the mean of these two was compared with nonblinded results by paired t test. Regression analysis was performed for the age of PD patients versus improvement in UPDRS and CAPIT scores 1 year after unilateral pallidotomy. Comparison of daily dosage of antiparkinson medications at baseline and 1-year follow-up was performed for both surgical and nonsurgical groups using ANOVA and paired t test. Results of all tests were determined to be statistically significant when the p value was less than 0.05.

Results. There was no perioperative morbidity or mortality in any of the patients. All surgical patients demonstrated improvement of rigidity and bradykinesia immediately after lesioning, while in the operating room. However, the first formal testing was obtained 1 week after surgery. The average length of hospitalization was 5 days, with all patients returning to their usual activities within 10 days.

Postoperative stereotactic MRIs (figure 2) confirmed the placement of all lesions within +-\1 mm of preoperatively planned targets. Lesion volume in all cases was calculated at between 60 and 90 mm³ with a mean of 75 mm³. Ten left and eight right pallidotomies were done.

There were only two events that could be considered complications. One patient was sexually disinhibited for 24 hours after pallidotomy, which was likely a transient effect of the surgery. The other patient suffered an MCA stroke 7 months after pallidotomy with infarction on the side opposite to the pallidotomy. He improved and became ambulatory within 2 to 3 months. As his 9- and 12-month follow-ups were confounded by signs of hemiparesis, they were excluded from the analysis; therefore, the 9- and 12-month follow-ups are reported for 17 patients.

For surgical patients, there was a significant decrease in UPDRS motor and ADL scores 12 HOM at 1 week and at 3, 6, 9, and 12 months postoperatively, which translates into significant clinical improvement Figure 3 and Figure 4.

CAPIT subtest scores were improved 12 HOM. Pronation-supination scores significantly improved contralaterally and ipsilaterally to the lesion at 1 week and 3, 6, 9, and 12 months postoperatively Figure 5and Figure 6. There was a significant improvement in contralateral finger dexterity and arm/hand movement between two points at all times tested postoperatively. Ipsilateral improvement was also observed. Walk scores improved significantly at 12 HOM at 12 months postoperatively (figure 7) in 10 patients who were able to complete both baseline and follow-up walk testing.

Blinded review of the randomized videotapes showed CAPIT results that did not significantly differ from the nonblinded initial measurements. All blinded results showed statistically significant differences between baseline and 12-month follow-up (table 1).

In general, UPDRS scores postoperatively improved significantly 12 HOM by a mean of 65%. CAPIT subtest scores postoperatively improved 12 HOM on the contralateral limb by a mean of 38.2%, and ipsilateral limb scores postoperatively improved by a mean of 24.2%. Walk scores improved 12 HOM by a mean of 45.0%.

Differences between "best on" baseline and 12-month follow-up were significant for all tests: UPDRS Motor and ADL, contralateral and ipsilateral upper extremity CAPIT, and walk CAPIT. However, those changes or improvements were less significant than those for 12 HOM.

There was no statistically significant improvement in UPDRS or CAPIT scores seen after initiation of study in nonsurgical subjects Figure 3^,Figure 4, and Figure 6.

At 1 year postoperatively, eight surgical patients demonstrated an increase in levodopa dosage or dopamine agonist use. Eight patients decreased levodopa usage or eliminated agonists, and one patient remained on the same medications and dosages. In contrast, five nonsurgical patients had increased dosage, one had decreased dosage, and one remained on the same dosage. The average levodopa dosage for surgical patients at the time of surgery was 1,041 mg/day and at the 12-month follow-up it was 1,002 mg/day, an insignificant 3.7% overall decrease. Pergolide dosage remained the same in the 10 patients who used it. Mean selegiline dosage was reduced to 50% of the baseline at 12-month follow-up in the 11 patients who used it, which is an insignificant difference despite the tendency for decrease (p = 0.08). Amantadine was eliminated in three of the four patients who used it preoperatively. Apart from one patient's stopping bromocriptine at the 12-month follow-up, no other changes were demonstrated in the three other patients who used it.

Nonsurgical patients increased levodopa usage from 740 mg/day at baseline to 977 mg/day at 12-month follow-up, a statistically significant 32% increase.

No correlation was found between the age of surgical patients and degree of improvement. Interestingly, one of the least improvements was seen in the youngest patient and one of the best improvements was seen in one of the oldest patients. If these two patients were excluded from the analysis, then younger age did correlate with better outcome in the remaining 15 patients.

Discussion. Our observations utilizing modern stereotactic and imaging technology not only corroborate previous reports ^[19,22,23] that ventral pallidotomy is an effective treatment for Parkinson's disease but more comprehensively delineate and quantify its areas of effectiveness. In 18 patients, we found a significant improvement in parkinsonian symptoms (tested at 12 HOM), including bradykinesia, rigidity, resting tremor, and ambulation difficulty, as well as a resolution of medication-induced contralateral dyskinesias.

Lesion volume in all our cases was between 60 and 90 mm³ with a mean of 75 mm³. Hariz ^[43] reported that the size of the pallidal lesions in his patients ranged from 28 to 150 mm³ with a mean of 67 mm³. Placement of lesions in all our patients was within 1 mm of the preoperatively planned targets. Hariz ^[44] found that the midlesion point in his patients was in a range from 4.25 mm medially to 1.25 mm laterally; 3.5 mm posterior, 4.25 mm anterior; 2.0 mm ventral to 3 mm dorsal. Although there is practically no difference in the mean lesion sizes, our results demonstrate less fluctuation in size and increased accuracy of localization when compared with the results of Hariz ^[43,44].

Our impression that younger PD patients showed greater improvement was not supported by statistical analysis. We still believe that there is a real tendency for this to occur, but due to a relatively small number of observations as well as a few exceptions, which were alluded to in the Results, no correlation was found. Preliminary analysis of a larger group of postpallidotomy follow-ups, although with shorter follow-up periods, shows that younger patients have greater improvement. However, as some older patients enjoy considerable improvement, age should not be a relative contraindication for pallidotomy; our oldest patient was seventy-nine.

Blinded review of all timed tests confirmed nonblinded clinical assessments; there was a significant improvement in PD symptoms and signs after pallidotomy. Patients experienced and reported improvements in overall functioning when in their "best on" and some decrease in fluctuation of their "on" and "off" periods; analysis of 12 HOM and "best on" improvements confirmed those observations. In comparing the degree of change or improvement for 12 HOM with "best on" 12 months after pallidotomy, 12 HOM showed significantly more improvement than "best on" for all observed parameters. Our patients were levodopa-responsive, and most had a good response to medication except for the presence of dyskinesias during short "on" periods. It was not expected that pallidotomy would dramatically improve a "best on" that was already good. Large improvements in 12 HOM, however, had to contribute to lesser fluctuations, as the "off" phase was better after pallidotomy and scores were lower, even approaching scores in the "on" phase.

Overall, there was a minimal decrease in the mean daily dosage of antiparkinsonian medications at 12 months postoperatively as compared with baseline. Postoperatively, the "barrier" of contralateral dyskinesia was lifted, allowing patients to tolerate the same dosage of antiparkinsonian medications, alleviating the parkinsonian symptoms without inducing contralateral dyskinesia.

Although pallidotomy may not be as effective as thalamotomy for tremor alleviation, tremor was reduced in all affected patients. Further, speech impairment, which is often a complication of thalamotomy, ^[45] did not occur following dominant or nondominant hemisphere pallidotomy. Also, whereas balance impairment can be a complication of thalamotomy, pallidotomy improved balance and ambulation as well as bradykinesia.

Dopaminergic hypoactivity in PD may result in a net increase of inhibitory neuronal output from the basal ganglia to the thalamus ^[46]. Depletion of striatal dopamine decreases activity of inhibitory striatal neurons. These striatal cells in turn project directly to the GPi (direct system), increasing the activity of inhibitory striatal neurons that project to the globus pallidus externus (GPe; indirect system) ^[47-49]. Excessive inhibition of GPe in PD potentiates subthalamic nucleus activity, which in turn causes hyperactivity of GPi. This hyperactivity causes excessive inhibition of the thalamus and thalamocortical fibers ^[50]. Improvement of bradykinesia after destruction of the subthalamic nucleus in MPTP-parkinsonian monkeys ^[26] provides experimental support for the efficacy of pallidotomy, since the primary output of the subthalamic nucleus is to the GPi ^[51]. Unlike the subthalamic nucleus, the GPi is surgically accessible in humans, with low mortality and morbidity. Rigidity may be improved by ventral pallidotomy, ansotomy (lesions in the ansa lenticularis), and combination of these techniques ^[52-58]. Ventral pallidotomy interrupts the majority of pallidofugal fibers passing through the globus pallidus ^[59-63]. Since pallidotomy was reported to improve parkinsonian symptoms long before dopaminergic therapy was available, it is unlikely that the effects of pallidotomy result from altering a condition related to chronic dopaminergic therapy as opposed to primary PD, but such a possibility must be considered as a contributing factor.

The mechanism of the ventral pallidotomy antiparkinsonian effect appears to be twofold: immediate and continuous. The immediate intraoperative effect may be to re-modulate the circuitry between the direct and indirect basal ganglia motor systems. The continuous or secondary effect may result from trophic influences, transsynaptic degeneration, or regenerative sprouting occurring over several months. These postulated mechanisms may also explain the ipsilateral improvement. The ipsilateral improvements with CAPIT testing and the UPDRS ADL and Motor scores, especially with regard to balance and gait, underscore the bilateral effects of unilateral basal ganglia lesions. Although other authors ^[64-67] have reported complications, including dementia, in patients undergoing bilateral pallidotomy, we have also begun performing second pallidotomies on the opposite side in selected patients, with encouraging early results.

Postpallidotomy improvement appeared immediately after surgery and persisted for the duration of the 1-year follow-up. Some additional but not significant improvements at 1 year suggest that pallidotomy might slow the progression of PD, because nonsurgical controls deteriorated. Such PD arrest was suggested ^[68] after some surgical interventions in the past; however, further follow-up is necessary to confirm this observation. It is conceivable that the presence of dysfunction and hyperinhibition can further perpetuate the imbalance and contribute to the vicious cycle of PD progression, but removal of GPi hyperactivity may reduce the possible contribution of neuronal circuitry hyperactivity-imbalance to the progression of PD.

We found that pallidotomy with short hospitalization produced a significant reduction in parkinsonism and contralateral "on" dyskinesia without morbidity or mortality in PD patients in whom medical therapy had failed. The improvement has continued throughout the course of this ongoing study. Similar to the most successful cases of fetal mesencephalic transplantation, ^[31] this procedure produces considerable improvement in "off" parkinsonism virtually immediately, without the complexities involved in fetal tissue use. Many subjects with severe bilateral parkinsonism have requested that a second pallidotomy be performed on the opposite side. Our findings, based on 1-year follow-up, suggest that stereotactic pallidotomy is a safe symptomatic therapy for selected Parkinson's disease patients.

Note. Readers can obtain 13 pages of supplementary material from the National Auxiliary Publications Service, c/o Microfiche Publications, P.O. Box 3513, Grand Central Station, New York, NY 10163-3513. Request document no. 05200. Remit with your order (not under separate cover), in US funds only, $7.75 for photocopies or $4.00 for microfiche. Outside the United States and Canada, add postage of $4.50 for the first 20 pages and $1.00 for each 10 pages of material thereafter, or $1.75 for the first microfiche and $.50 for each fiche thereafter. There is a $15.00 invoicing charge on all orders filled before payment.

Acknowledgments

We are indebted to Dr. Lauri Laitinen for his generous advice and encouragement; to Drs. Wen C. Yang and David Liu for neuroradiologic assistance; to the late Dale Samelson, RN, and to Linda Chin, RN, and Sheree Loftus, MSN, CRRN, for important clinical and research contributions. Thanks are due to Drs. Alan Kluger and Kiril Kiprovski for their expert assistance in statistical analysis.