Ventroposterior pallidotomy, and pallidal and subthalamic nucleus (STN) stimulation, are proposed to treat severe Parkinson’s disease (PD) with levodopa-induced motor complications. Pallidal surgery decreases levodopa-induced dyskinesias (LID). Parkinsonism evaluated by the motor part of the Unified Parkinson’s Disease Rating Scale (UPDRS) improves by about 30% 1, 2 after unilateral pallidotomy. STN stimulation can greatly improve parkinsonian disability, but its effects on LID are unknown3 because the first patients treated with STN stimulation were selected with mild or absent LID since spontaneous STN lesions can induce hemiballism. However, our initial experience with STN stimulation showed that stimulation-induced abnormal movements could be controlled by setting the stimulation voltage below the dyskinesia threshold. Moreover, the substantial decrease in levodopa dosage, which was made possible by the antiparkinsonian benefit, allowed a decrease in LID in the long-term. Thus, STN stimulation was also proposed for patients with severe LID. We selected all consecutive patients with young-onset PD, since this population is generally levodopa-sensitive and prone to develop severe LID. Eight patients with a mean age of 51 (SD 10) years and a mean disease duration of 16 (5) years were operated on bilaterally in the STN because of severe motor fluctuations (motor score on the UPDRS 57· 5 [14· 5] off-drug and 18· 2 [10· 6] on-drug). The study was approved by the Grenoble University ethical committee and all patients gave their written, informed consent. The surgical procedure has been described previously. 3 6 months after surgery, the motor score of the UPDRS improved from 58 (16) to 17 (8)(–71%) in the off-drug condition when stimulation was switched on. This improvement was seen in all the major parkinsonian symptoms. The dyskinesia score (six body parts, each scored 0–4, maximum score 243), assessed during the maximum motor response of a levodopa test with the same dose as preoperatively (early morning+ 50–100 mg), decreased from 11· 4 (6· 6) preoperatively to 6· 8 (4· 3) postoperatively onstimulation. Dyskinesia duration and disability (UPDRS items 32+ 33), which assess LID during activities of daily living under chronic stimulation, decreased from 4· 0 (1· 9) to 1· 4 (0· 9)(–65%). The antiparkinsonian drug dosage calculated as levodopa equivalent1 decreased from 1560 (930) to 680 (420) mg daily (–56%).

With chronic STN stimulation, LID were greatly improved, in a similar percentage as with bilateral globus-pallidusinternus stimulation. 4 With unilateral pallidotomy, 1, 2 contralateral LID are substantially improved although a big reduction of levodopa is not possible. Targeting the STN instead of the globus pallidus internus has the advantage of inducing a more pronounced antiakinetic effect. 1–4 The decrease in LID induced by STN stimulation can be explained mainly by the decrease in levodopa dosage. This is why the improvement in dyskinesias is more obvious in activities of daily living than during the levodopa test with a dose identical to that given before surgery. Moreover, as highfrequency stimulation is supposed to act through neuroinhibition, the chronic inhibition of STN neurons can change the dyskinesia threshold with time, in the same way as the progressive abatement of hemiballism after a vascular lesion of the STN. 5 Replacing the pulsatile levodopa-induced decrease of STN activity by a chronic stable decrease seems to improve dyskinesias. After 4 years of experience of STN stimulation, 3