Assessment of motor function after stereotactic pallidotomy

Stereotactic surgical lesioning of the globus pallidus internus (GPi) is gaining acceptance as a treatment for advanced Parkinson's disease(PD).^1-8 The procedure was pioneered in the 1950s but was eclipsed by the development of surgical thalamotomy and the advent of levodopa therapy.^9-14 More recently, however, improved understanding of basal ganglia circuitry, advances in neuroimaging and physiologic recordings, and a full appreciation for the complications of chronic levodopa therapy have rekindled interest in the procedure.

At the present time, despite increasing popularity of pallidotomy, there are few controlled data about its efficacy, and many questions remain: identification of ideal candidates, specific target sites and methods for localization, the duration of benefit, and the determination of clinical characteristics that predict improvement.^1,15-18 Most pallidotomy series have reported robust improvement in contralateral levodopa-induced dyskinesia,^1-3,7,19-25 and it has been suggested that pallidotomy produces most of its subjective benefit by merely reducing levodopa complications rather than by improving the underlying PD. The effects on bradykinesia, tremor, gait, postural instability, and speech have been less well established. The assessment of these measures when evaluated in the "off" state, without the benefit of levodopa, should best determine the procedure's effect on the underlying motor dysfunction.

The Unified Parkinson's Disease Rating Scale (UPDRS) provides a quantitative assessment of motor function in PD.²⁶ We evaluated randomized videotapes of 34 patients before and 3 months after unilateral pallidotomy in order to determine the effect of the procedure on motor function in the "off" state.

Methods. Patients were recruited from the Baylor College of Medicine Parkinson's Disease Center and Movement Disorders Clinic. All patients satisfied the following inclusion criteria: (1) Hoehn and Yahr stage 3 or greater while "off," (2) levodopa responsiveness and a history of levodopa-induced dyskinesias, and (3) absence of other neurologic disease including significant dementia or autonomic dysfunction. Data from the first 34 consecutive patients completing 3-month evaluations were included in the study. Patients underwent standardized videotaping before surgery and 3 months after surgery as part of the modified Core Assessment Program for Intracerebral Transplantations (CAPIT) protocol.²⁷ The video recording included assessments of writing, a standardized passage reading, and all aspects of the UPDRS motor examination except hand pronation/supination, and postural stability. Rigidity was also omitted because this could not be assessed by videotape. Video recordings were made during the "off" state at least 12 hours after the last ingestion of levodopa. Subjects wore caps to avoid unblinding the observers (W.O., C.S.). Tapes were edited into randomly sequenced segments and blindly rated by two observers at separate times.

The surgical procedure is described in detail elsewhere.²⁸ To briefly summarize, after fixation of the Leksell G frame, stereotactic contiguous 1-mm axial scans were obtained via a helical CT scanner. An initial target in the posteroventral GPi was chosen 20 mm lateral to the midline, 4 mm below the anterior commissure-posterior commissure (AC-PC) line, and 2 mm anterior to the midcommissural point. Reformatting the images in axial, coronal, and sagittal planes as well as comparison with preoperative MRI was used to ensure accurate identification of the commissures and the target. A burr hole was made 7 to 9 cm above the orbital rim and 2 cm lateral to the midline. Microelectrode trajectories were planned at an angle of 35° to 45° above the AC-PC line and parallel to the midsagittal line. Microelectrodes were used that had an initial impedance measured between 0.5 and 1.2µΩ. The microelectrode was advanced via the guiding cannula, and single-unit recordings were obtained from 20 mm above the tentative target to 5 mm below it. The ventral border of the GPi was identified with the recording. The optic tract was identified in many cases with microstimulation. One to three trajectories were made in most cases, and the microelectrode recording usually resulted in modifying the CT coordinates by 1 to 3 mm, most often in the vertical (Z) axis. We used a 1.1-mm unipolar electrode with a 3-mm uninsulated tip for macrostimulation and lesion-making. The tip of the electrode was placed within 1 mm of the ventral border of the GPi according to the mapping. Macrostimulation was used to assess the threshold for the spread of current to the optic tract and the internal capsule. This location was considered the final target in cases in which no optic or capsular responses were elicited below 2 volts. Otherwise the electrode was withdrawn in 1.0-mm increments, and the stimulation was repeated. A radio-frequency lesion was made at 75 °C for 60 seconds. The electrode was then withdrawn 2 mm, and a second similar lesion was made. During lesioning, the strength and mobility of the contralateral arm, speech, and visual fields were monitored. Total operative time averaged between 3 and 4 hours, and length of hospital stay averaged 4 days.

Inter-rater reliability was determined by the Wilcoxon's sum test and rank correlation test. Paired Wilcoxon's signed rank testing was used for statistical analysis of postoperative results, and factors correlating with improvement and lack of improvement were determined using unpaired t-tests.

Results. Thirty-four patients (18 female) with PD underwent unilateral pallidotomy and blinded assessments before and 3 months after the procedure. Their mean age at surgery was 59.2 ± 9.1 years (range, 40 to 74 years), and all were right-handed. The mean age at initial symptom onset was 45.0 ± 9.1 years (range, 24 to 61 years), resulting in a symptom duration of 14.2 ± 5.6 years (range, 4 to 25 years). Because dysfunction tended to be worse on the right side, and patients usually desired improvement in their dominant hand, we performed 25 procedures on the left brain and 9 on the right brain.

Five patients reported mild and transient postoperative adverse events. One had slight Broca's aphasia that resolved by 2 months, and four demonstrated postoperative altered mental status that had resolved by the time of hospital discharge. Postoperative MRI demonstrated a 9-to 10-mm lesion and surrounding edema in all patients. Six months after the procedure, MRI in 32 patients demonstrated lesions in 29. Mean lesion size was reduced to 22.0 ± 28.8 mm³ (range, 0.5 to 94 mm³). Twenty-three of these (79.0%) were exclusively within the GPi, 2 (6.9%) were in the GPi and globus pallidus externus (GPe), 3 (10.3%) were in the GPi extending into the internal capsule, and 1 (3.4%) was exclusively within the GPe.

At the preoperative assessment, the total UPDRS "off" motor score(determined by averaging the scores of two independent raters) was 28.9± 7.5. This improved by 13.6% to 25.0 ± 7.0 at the 3-month postoperative assessment (p ≤ 0.001). The total UPDRS "off" scores improved in 27 of the 34 patients (p ≤ 0.0005, binomial test for proportions). Seven patients (20.6%) did not demonstrate any improvement in "off" scores, and two others (5.9%) improved by less than 1 point. When compared with good responders, this group of nine poor responders was of similar age (58.8 ± 10.1 versus 58.7 ± 8.7) and symptom duration (12.8 ± 4.7 versus 14.7 ± 5.8), but did have lower total preoperative UPDRS scores (22.7 ± 9.6 versus 31.0 ± 5.5, p = 0.028). In one poor responder, postoperative MRI demonstrated the lesion in the GPe rather than the GPi. Another patient whose total UPDRS score improved by 2.3 points subsequently developed features of atypical parkinsonism and was clinically thought to have multiple system atrophy. No other patient demographic or surgical variables correlated with poor outcome.

Nineteen total data points were assessed on each patient by two raters, and the mean score was used for statistical analysis. Inter-rater reliability was generally good. Rank correlation of the preoperative to postoperative difference was statistically significant (p < 0.05) in 13 of 19 categories (68.4%), and difference scores of the two raters were not significant in any variable using Wilcoxon's rank sum test.

Eighteen of 19 categories (p ≤ 0.00005) improved, 11 of these (57.9%) to a statistically significant degree (p < 0.05)(table). The most robust improvements were seen in contralateral tremor, gait, and arising from a chair. Ipsilateral tremor, contralateral dexterity, ipsilateral heel tapping, and body bradykinesia also improved significantly. All other measures, except for facial expression, as discussed below, tended toward improvement. All subsections of the motor UPDRS improved significantly (figure). Total tremor improved by 59.5%, gait scores improved by 21.9%, and total dexterity scores improved by 12.5%. Although various contralateral measurements tended to be superior to those ipsilateral to the lesion site, only rest tremor demonstrated statistically significant superiority when the two sides were compared against each other (p ≤ 0.01).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Mean averaged UPDRS "off" motor scores plus 95% confidence intervals before and after pallidotomy. *p ≤ 0.05, **p≤ 0.01, ‡p ≤ 0.001, ‡‡p ≤ 0.001.

Discussion. Our study, based on blinded review of videotapes, demonstrates objective improvement in the cardinal signs of PD at 3 months after palidotomy that is independent of the effects of levodopa. Almost all features of the disease improved, especially measures of contralateral tremor, gait, and posture. There was also significant, but less robust, improvement ipsilateral to the side of pallidotomy. No factors emerged that reliably predicted poor surgical outcome other than suboptimal lesion site and relatively lower initial UPDRS scores. We suspect that surgery on patients with less severe parkinsonian signs does not allow for as dramatic improvement secondary to simple ceiling effect. However, some underlying pathophysiologic variable, present in more advanced clinical cases, may predispose those patients to a more dramatic improvement.

Previous assessments of post-pallidotomy "off" motor function have varied from no improvement to fairly significant improvement.^20-24,26 Our findings, however, generally concur with the majority of smaller and unblinded series that show moderate "off" time benefit.^20-24 In an open, unblinded study of 15 patients, Baron et al.²⁰ reported a 24.9% improvement (p < 0.001) in the "off" UPDRS motor examination score at 6 months. UPDRS "off" gait, postural stability, and ipsilateral tremor improved at 3 months, while CAPIT timed measurements of akinesia improved bilaterally. In an unblinded evaluation of 18 PD patients, Dogali et al.²² reported a 21% reduction in "off" motor scores at 3 months that was maintained at 1 year. Total UPDRS scores improved 65%, and CAPIT timed tests of akinesia were statistically improved bilaterally. UPDRS subsets were not mentioned. Lozano et al.²¹ studied 14 patients in a manner similar to ours. Based on a blinded review a randomized videotapes, they reported a 30% reduction in total "off" UPDRS scores and statistically significant improvement in bilateral akinesia scores (UPDRS items 23 to 26) and contralateral tremor (items 20 and 21). Postural stability and gait score improvement did not reach statistical significance, although combined postural instability and gait disorder CAPIT scores did improve significantly. They concluded that pallidotomy immediately improved "off" period symptoms of PD, but aside from reduced contralateral dyskinesia, did not significantly augment "on" time functioning.

Although the pattern of motor improvement is generally consistent among reports, degrees of improvement do vary. Variables that may account for these differences include lack of blinding in some series, sensitivity of the rating scales and raters, patient population, and several surgical techniques including lesion location, lesion size, and localization methods. For example, the report by Lozano et al.²¹ tended to show superior upper-extremity improvement, but less dramatic lower-extremity improvement when compared with our patients, possibly suggesting different somatotopic representation of lesion sites. Determination of the optimal surgical variables awaits careful comparisons among centers currently performing pallidotomy.

The improvement in "off" scores seen on the contralateral side of surgery is consistent with traditional models of the basal ganglia physiology in PD that postulate tonic thalamic hypoactivity resulting from excessive inhibition from the GPi.²⁹ The GPi efferents somatotopically connect to the ventral lateral thalamus via the ansa lenticularis (anteriorly) and fasciculus lenticularis (posteriorly). There are, however, other less well-defined GPi efferents that may play a role in motor function. The excessive GPi activity is the result of reduced striatal inhibition (direct pathway) and increased sub-thalamic excitation (indirect pathway). This model does not account for improvement ipsilaterally to the side of the lesion. One recent study, however, that compared preoperative and postoperative 18F-fluorodeoxyglucose PET scans in eight patients who underwent unilateral pallidotomy demonstrated bilateral (ipsilateral more than contralateral) increases in premotor, supplementary motor, and dorsilateral prefrontal cortical activity at rest.³⁰ Several other anatomic and physiologic studies support the presence of bilateral cortico-striatal-thalamic inter-hemispheric connections; however, a definitive neuroanatomic substrate for this communication has not been clearly elucidated.^31-33 The relatively more pronounced improvement in contralateral tremor, and levodopa induced dyskinesia, suggest that separate pathways may subserve these motor features.

Motor function is only one aspect of PD that can be affected by pallidotomy. Other features, including motor function while taking dopaminergic therapy, effect on dyskinesia, cognitive responses, autonomic dysfunction, and subjective reports, were not evaluated in this report. Because most patients continue to take dopaminergic therapy after surgery, the improvement in "off" time motor function may not represent the actual benefit received by the patient. Nevertheless, we used the "off" functioning as a means of assessing the procedure's effect on underlying motor deficits in PD without the contaminating effect of levodopa. Other limitations of this study include the relatively short follow-up time, the necessity of rating via videotape rather than by personal examination, and the effect of the long duration response of levodopa. We were unable to rate rigidity, and intermittent rest tremor may have been missed during the 10 to 15 minutes of each video segment. It is also our clinical impression that hypomimia (masked facies) does improve, but that the formal nature of the video recording limited demonstration of facial expression.

Our study shows that pallidotomy improves the cardinal motor signs of PD; however, longer follow-up time is required to assess its long-term effects. The effect of the procedure on rigidity and other features, including levodopa dyskinesias and cognitive function, will be reported separately. Although improvement in dyskinesia further augments the overall patient satisfaction after pallidotomy, the absence of this symptom should not deter consideration of the procedure in otherwise advanced PD.