A study of medial pallidotomy for Parkinson's disease: clinical outcome, MRI location and complications.

We have studied the effects of unilateral ventral medial pallidotomy in 26 patients with medically intractable Parkinson's disease with marked drug-induced dyskinesias. Preoperatively, all patients were assessed during one 5-day admission according to the Core Assessment Programme for Intracerebral Transplantation (CAPIT) protocol, including rating in the 'practically defined off' and 'best on' states before and during a single-dose levodopa challenge. Motor performance was assessed with subset categories of the Unified Parkinson's Disease Rating Scale (UPDRS), timed motor tests and a standard dyskinesia rating scale. Pallidotomy was performed under stereotaxic CT guidance with intra-operative extracellular microelectrode recording made from the basal ganglia. All patients were re-assessed 3 months postoperatively and a subgroup (n = 9) have so far also been re-assessed after 1 year. Pre- and postoperative performance scores were compared in order to determine which categories of performance improved postoperatively. Significance was accepted at P < 0.005 in order to take into account the multiple number of comparisons performed. Patient medication was compared pre- and postoperatively and the morbidity associated with surgery was also recorded. The most significant improvement postoperatively was the diminution of 'on' dyskinesias contralaterally (67%, P = 0.0001); however, ipsilateral (45%, P = 0.0006) and axial (50%, P = 0.0008) dyskinesias also improved. Contralateral to pallidotomy, the median 'off' motor UPDRS score improved by 27% (P = 0.001) and a significant improvement was also observed in contralateral rigidity by 25% (P = 0.001). There were trends towards improvement in contralateral tremor (33%, P = 0.016) and bradykinesia (24%, P = 0.013) scores. Ipsilateral rigidity improved by 22% (P = 0.005), but other ipsilateral motor scores did not alter significantly. The 'off' gait/postural instability score and 'off' walking time showed marginally significant improvements by 7% (P = 0.007) and 29% (P = 0.014), respectively. On medication, no significant postoperative improvements in parkinsonism were detected. Anti-parkinsonian medication increased by 11% postoperatively. In the subgroup who were available for assessment 1 year postoperatively, responses were generally maintained. Two (7.7%) of the 26 patients had fatal complications (one cerebral haemorrhage and one haemorrhagic infarct) directly related to surgery. Among the remaining 24 patients, four (15.4% of the total 26) had major complications (two persisting and two transient). Ten patients (38.5%) had minor complications. The majority of the complications (major and minor) occurred in the earlier operated patients and the complication rate subsequently declined with increasing operative experience. The remaining 10 patients (38.5%) had no significant side-effects. One of these 10 patients died from an incidental malignant glioma 6 months postoperatively. These findings confirm that levodopa-induced dyskinesias are dramatically reduced following ventral medial pallidotomy and constitute the principal indication for pallidotomy. Improvements in underlying parkinsonism were of smaller magnitude. Pallidotomy may also offer some patients an opportunity to increase antiparkinsonian medication. Patient selection for medial pallidotomy should, therefore, be based largely on anticipated improvements in levodopa-induced dyskinesias, but this must be balanced against the associated morbidity and mortality.