Deep brain stimulation (DBS) in the thalamic ventral intermediate nucleus (Vim) was introduced by Benabid et al as an alternative to thalamotomy for the treatment of tremor of parkinsonian or other origin.1 It soon appeared that the advantages of Vim DBS over thalamotomy were the non-ablative nature of this surgery, the possibility of bilateral surgery, the reversibility of possible stimulation related neurological side effects and the adaptability of stimulation (up to a certain limit) according to the needs of the patient.2^–5 Although Vim DBS has been widely used in the treatment of parkinsonian tremor, there are few and contradictory reports on the results of this surgery beyond 1 year of follow-up.6^–9 Only three studies to date have reported long term outcome after 5–6 years.10^–12

In 1999, Limousin et al published the 1 year results of patients with Parkinson’s disease (PD) or essential tremor (ET) who had Vim DBS in the framework of the European multicentre study.13 The results at 6 years for patients with ET in this study have been published, confirming the long term efficacy of Vim DBS in ET.14 Here, we evaluated patients with PD from this study at a mean of 6.6 years after surgery. The primary aim was to document the long term effect of Vim DBS on contralateral tremor of PD. Additionally, the long term effects of Vim DBS were assessed on other parkinsonian symptoms and on activities of daily living (ADL), allowing appraisal of the value of Vim DBS in a population of highly selected patients with tremor as the main disabling symptom.

PATIENTS AND METHODS

Eight of the original 13 centres,13 with a total of 66 patients with PD, participated in this long term follow-up. Of these 66 patients, 19 had died of causes unrelated to the procedure and seven patients were lost to follow-up. Additionally, one patient refused follow-up and one patient was demented and could not be evaluated. Thus the long term study was performed in a total of 38 patients. Table 1 shows the characteristics of the patients at baseline and at follow-up.

Inclusion and exclusion criteria of patients enrolled in the European multicentre study have been documented in the 1 year follow-up publication.13 The initial study was approved by the ethics committees of the centres involved. For this long term follow-up, patients gave their written informed consent to participate. In this paper, we present the results at 1 year and at maximal follow-up (ie, minimum of 6 years after surgery).

Surgery was performed without microelectrode recording in seven centres and with microelectrode recording in one. Stereotactic ventriculography and/or stereotactic CT or MRI were used for target determination. The target was the Vim nucleus of the thalamus. The electrical parameters of chronic stimulation (polarity of the quadripolar electrode, amplitude of current, pulse width, frequency) were adjusted as needed after surgery to obtain optimal effect on tremor with as few side effects as possible.

This long term study was designed to evaluate the additional effect of thalamic DBS on tremor in patients taking their regular antiparkinsonian medication. Therefore, all evaluations were done while patients were taking their regular medication. All assessments were done in the morning, after the patients had taken their regular morning dose. Scoring of the motor items of the Unified Parkinson’s Disease Rating Scale (UPDRS) part III was performed first under “off stimulation” then under “on stimulation” conditions. The ADL part of the UPDRS (part II) was also evaluated. We also recorded the intensity and duration of dyskinesias when present, and the proportion of “off” periods per day (items 32, 33 and 39 of UPDRS part IV). We assessed Hoehn and Yahr staging, and Schwab and England’s ADL scale under “on drug, on stimulation” conditions.

The following subscores of the UPDRS part III were analysed separately for treated and non-treated body side (ie, contralateral and ipsilateral to the stimulated thalamus): tremor (items 20 and 21), rigidity (item 22) and akinesia (items 23, 24, 25, 26). In addition, subscores of UPDRS III dealing with speech (item 18), gait (item 29) and postural instability (item 30) were evaluated in the group as a whole, as well as separately in those patients who had unilateral or bilateral thalamic DBS, respectively. From part II of the UPDRS, the following individual items were analysed: handwriting (item 8), cutting food (item 9) and dressing (item 10). Changes in doses of anti-parkinsonian medication were compared between baseline and the latest follow-up. Changes in stimulation parameters were compared between the 1 year and latest follow-up. Finally, adverse events that could be attributed to the stimulation, including the need to replace the stimulator because of battery failure, were recorded at the 6 year follow-up.

Data are presented as mean (SD) and range, preoperatively, at 1 year and at maximal follow-up. To evaluate the time effect of stimulation, we compared the results at 1 year and at maximal follow-up (mean 6.6 years) with those at baseline. To evaluate the effect of therapy, we compared the results at 1 year as well as those at maximal follow-up with stimulation on compared with stimulation off. For statistical analysis, the Wilcoxon’s test for paired samples was used. A p value of 0.005 was considered to indicate statistical significance with use of the Bonferoni correction method.

RESULTS

At baseline and at 1 year, there were 32 unilaterally operated and six bilaterally operated patients. Between 1 and 6 years, two of the patients received an additional contralateral thalamic DBS. Thus at 6 years there were 46 treated body sides in the 38 patients. Mean follow-up was 6.6 (SD 0.48) years (range 5.9–7.8). Patients exhibited no worsening of off periods or dyskinesias, either in duration or in severity at maximal follow-up compared with baseline (table 1). Medication doses remained stable but disease staging according to Hoehn and Yahr showed significant worsening at 6 years (table 1).

Tremor scores

Table 2 depicts the UPDRS scores for tremor on the treated side and contralaterally. On the treated body side, tremor was significantly less severe at 6 years than at baseline when stimulation was on, and also when stimulation was off. Additionally, tremor scores at 6 years were further improved when stimulation was on compared with when it was off. Tremor scores on the non-treated side, whether on or off stimulation, also tended to decrease over time, mainly between year 1 and year 6, but this change was not significant.

Total motor scores

Table 3 shows that the scores of the motor part of the UPDRS at 6 years with stimulation off were similar to those at baseline, but they were better when stimulation was on than when it was off.

Appendicular akinesia and rigidity

Table 4 shows a minor but significant improvement in appendicular akinesia and rigidity items on the treated body side when stimulation was on compared with when it was off and compared with baseline. There were no differences for akinesia and rigidity on the non-treated side between scores at 6 years compared with scores at baseline.

Axial scores

Table 5 shows that “axial” motor symptoms (speech, gait and postural stability) deteriorated markedly at 6 years compared with baseline, both under the on and off stimulation conditions, and this change occurred over time between year 1 and year 6. The deterioration of speech occurred equally in patients with unilateral and bilateral DBS (results not shown in table 5).

Functional disability

Table 6 shows a significant improvement in functional disability at 1 year according to Schwab and England’s ADL scales as well as items of the UPDRS part II related to handwriting, cutting of food and dressing. None of these improvements, however, persisted at the 6 year follow-up despite maintained improvement in tremor scores.

Parameters of stimulation

While voltage and frequency, but not pulse width, had to be increased during the first year, at 6 years there were no differences in these parameters compared with year 1. Amplitude at 6 years was 2.4 (0.8) V (range 0.5–3.6). Frequency was 160.7 (30.2) Hz (range 130–185). Pulse width was 79.3 (28.3) μs (range 60–180). At 6 years, 26 patients were stimulated monopolarly using one lead contact, 11 using two adjacent contacts and one patient using three adjacent contacts. None had bipolar stimulation.

Adverse events

There were a total of 43 adverse events in the 38 patients. Adverse events related to the device included 17 replacements of the stimulator because the batteries expired, one repositioning of the stimulator, one repositioning of the DBS lead, three infections and one skin erosion without infection. Adverse events reported by the treating centres as being related to stimulation included two cases of dystonia, two dysarthria, three dysequilibrium, one hallucination and one dementia. There were 11 other “non-classified” adverse events (gait disorder, paresthesia, paresis, dyskinesias, proximal limb ataxia, migraine, neck pain). Some patients had more than one adverse event.

DISCUSSION

At 1 year and at a mean of 6.6 years after surgery, tremor was well controlled by Vim DBS. Stimulation parameters were stable between 1 year and maximal follow-up, suggesting absence of “tolerance” to the treatment. There was a definite but slight disease progression in this group of patients (tables 4, 5), mainly of axial symptoms (speech, gait and postural stability), also reflected by an increase in Hoehn and Yahr staging. At 1 year, there was a functional benefit on ADL which was influenced by tremor for such things as handwriting, cutting food and dressing. However, this functional benefit was not maintained in the long term (table 6), despite ongoing tremor control, reflecting progression of other symptoms with ongoing disease. As tremor was well controlled (tables 2, 3) in this highly selected patient group with non-fluctuating tremor dominant disease, there was no requirement for an increase in dopaminergic medication and these patients did not develop or have worsened motor fluctuations or dyskinesias, as would have been expected after such a long observation period.

Although the study was prospective, there was no control group of non-operated patients or of patients operated on in other brain targets. The present study was designed after the end of the 1 year follow-up study,13 to obtain data on long term follow-up of Vim DBS. In the initial study, patients were assessed both on and off drugs and only the off drug results have been published. As was the case in the initial study, only a minority of patients were fluctuating (9/38 patients in the present study). It was therefore decided to evaluate patients while they were on their regular medication at the long term follow-up and to compare these data with the baseline on drug evaluation. This design had the advantage of better feasibility. There was no worsening of motor complications and the data therefore truly reflect the condition of the patients in everyday life.

In the long term, tremor, which had been the main disabling symptom in this selected patient group, was well controlled with Vim DBS. Interestingly, on the treated side, even when stimulation was off, there was more than a 50% decrease in tremor scores at 6 years compared with baseline. This phenomenon has also been reported by the Toronto group who thought that chronic DBS might induce permanent physiological changes in some patients leading to improvement in tremor.7 The same phenomenon, albeit not statistically significant, occurred on the non-treated side in our patients, indicating that the decrease in tremor over time is rather related to the natural course of the disease than to chronic VIM stimulation.

The fact that akinesia and rigidity showed a modest albeit significant amelioration when stimulation was on, both at 1 year and at 6 years compared with baseline, may reflect the well known fact that these symptoms may be more easily and perhaps more accurately scored when tremor is relieved. Indeed, the 4–6 Hz parkinsonian tremor interferes with the repetitive movements of the UPDRS. Moreover, the presence of a tremor can activate rigidity. Hence, given the very modest improvement of akinesia and rigidity, thalamic DBS does not seem to ameliorate rigidity or akinesia to a meaningful extent and the scored amelioration may rather reflect an evaluation artefact.

The total motor UPDRS scores were unchanged at 6 years compared with baseline. Hence, apparently there was no progression of disease in these patients. However, while motor scores related to tremor decreased, the motor scores related to speech, postural stability and gait markedly worsened. This is in accordance with the progression of the disease on chronic levodopa treatment15 or on chronic subthalamic stimulation.16 Thalamic stimulation can induce worsening of speech in individual patients, as shown for the side effects, and so an eventual role of thalamic DBS, especially bilateral, in this worsening cannot be ruled out. However, speech worsened significantly at 6 years, but not at 1 year, regardless of stimulation being off or on, and irrespective of patients being unilaterally or bilaterally operated. Worsening of speech in the whole group of patients thus seems to be mainly related to progression of the disease.

The ADL scores according to Schwab and England had improved at 1 year but returned to baseline at 6 years. As UPDRS ADL (UPDRS part II) contains a mixture of items related to PD symptoms and those related to ADL proper,17 we scored selected items of this scale directly related to ADL, such as handwriting, cutting food and dressing. These items had improved at 1 year but also returned to baseline at 6 years. Hence the definite decrease in tremor and apparent stability of contralateral appendicular motor symptoms at 6 years were not enough to maintain the improvement in ADL scores. In essential tremor, both short term and long term functional outcome is much better as in this disease tremor is the only symptom with an impact on ADL and progression of the disease does not annihilate the functional benefit of tremor improvement.13 14

Which is the preferable target for parkinsonian tremor, Vim or STN?

Our patients had a tremor dominant PD, and it is well known that patients with tremor dominant PD can have a relatively benign course of their disease.18 19 In these patients, tremor remained improved and there was only mild deterioration in other on drug motor symptoms at the 6 year follow-up compared with baseline, mainly affecting axial symptoms (speech, gait, postural instability) compatible with the natural history of the disease.15 18 20^–22 Disease evolution in these highly selected patients is obviously quite different from that in the general PD population.23 Tremor as the main symptom remained well controlled and patients did not require an increase in dopaminergic medication. This may well have contributed to the fact that in our patients as a group, neither motor fluctuations nor dyskinesias worsened during 6 years on chronic Vim DBS.

On the other hand, in 65% of the patients with Vim DBS followed by the Grenoble group for more than 4 years, motor fluctuations and/or dyskinesias were the most prominent complaints.24 Subthalamic nucleus (STN) stimulation has been shown to improve ADL in patients with previous Vim surgery who subsequently developed disabling motor complications of dopaminergic treatment leading to a second surgical intervention after a mean delay of 10 years.25 Aside from the positive effect of bilateral STN DBS on tremor,24 its effect on functional disability in fluctuating patients is more prominent and remains more stable in the long run16 than that of unilateral Vim DBS. Therefore, STN stimulation should be preferred over Vim stimulation even in patients with tremor dominant PD. However, subthalamic stimulation is more invasive and is more difficult to manage compared with Vim surgery: it generally has to be performed bilaterally, requires major adaptations of dopaminergic treatment and can lead to behavioural complications.26 Therefore, in elderly non-fluctuating patients with a stable mostly unilaterally dominant tremor as the main symptom, perhaps the more lenient thalamic Vim DBS may be suggested.