How do parkinsonian signs return after discontinuation of subthalamic DBS?

Patients and methods.

We studied 35 patients with advanced PD (age 64.2 ± 8.0 [mean ± SD], duration of disease 14.4 ± 4.9 years) with severe untreatable motor fluctuations. All patients had 1) a dopa-responsive parkinsonism (at least 25% improve-ment of preoperative Unified Parkinson’s Disease Rating Scale [UPDRS] motor score after l-dopa and apomorphine challenge), 2) no atypical red flag signs on neurologic examination,12 and 3) Mini-Mental State Examination (MMSE) score >24 and no dementia as defined by the International Classification of Diseases (ICD-10).

All patients had been bilaterally implanted 6.7 ± 3.3 months before in the STN by MRI and electrophysiologically guided stereotaxic neurosurgery with a quadripolar electrode (3389; Medtronic, MN) and an implantable pulse generator (IPG/Medtronic Itrel II or Kinetra). Their stimulation parameters were stable since 4.3 ± 2.7 months. On the right side, the stimulation was as follows: amplitude, 2.5 ± 0.7 mV; pulse width, 70 ± 21 μsec; frequency, 149 ± 38 Hz. For the left: amplitude, 2.6 ± 0.7 mV; pulse width, 74 ± 23 μsec; frequency, 151 ± 38 Hz. Presurgical Activities of Daily Living (ADL) “on” at the time of examination was 10.8 ± 4.7, similar to the preoperative “on medication” value of 10.8 ± 4.7. Hoehn and Yahr stages were unchanged at 2.5 ± 0.6 compared to 2.5 ± 0.7. Presurgical medication was 1,220 ± 510 mg levodopa equivalent dosage (LED)13; postsurgical, 115 ± 205 mg LED. Twenty-four patients were free of all antiparkinsonian medication for more than 2 months at the time of evaluation, 22 of them since their implantation (6.6 ± 3.9 months). Eleven patients were still using medication (69 ± 16% postoperative reduction of LED). The medication was stable since 3.7 ± 2.3 months.

Baseline evaluation was achieved with DBS switched “on” without medication (practically defined medication off state, following CAPSIT guidelines). DBS was then switched “off.” Clinical evaluations of motor UPDRS scores were performed at baseline and at 5, 15, 30, 60, 90, 120, 150, 180, and 240 minutes after switching DBS “off.” DBS was then switched back “on,” and clinical evaluations were performed after 5, 15, 30, 60, 90, and 120 minutes.

All patients gave their informed consent. The study was approved by the Ethical Committee of the Department of Medicine, CHUV.

Five of the 35 patients expressed discomfort during the test, so that time of DBS “off” observation had to be reduced to 2 hours; these results were excluded from the study. The remaining 30 patients were all studied for 3 hours. In addition, 11 of these patients agreed to be studied with DBS off for 4 hours, and 3 for 6 hours. All patients agreed to be studied sequentially with DBS “on” for 1 hour before leaving: 10 of them were studied over 2 hours. All examinations were performed by the same examiners (P.T., J.G., F.J.G.V.), assessed with video films, and coevaluated with the same senior neurologists (J.G., F.J.G.V.). In case of disagreement, the scoring of the senior neurologist was retained for the analysis.

We analyzed changes in the mean UPDRS motor score14 (maximal possible score 108). We calculated subscores for bradykinesia (sum of subitems 18, 19, 23 through 26, 31; maximal possible score 44), tremor (20, 21; maximal possible score 28), rigidity (22; maximal possible score 20), and axial signs (27 through 30; maximal possible score 16). In addition, worsening or improvement at each time of observation were also expressed as a percentage of the maximal change observed during the whole evaluation. TI (or TD) 75 and TI (or TD) 90 values are the times at which the mean group responses reached 75% and 90% of the maximal mean group change (TI 75 and TI 90 for increase; TD 75 and TD 90 for decrease). Comparisons were performed using Student t-tests; Bonferroni correction was applied to adjust the significance level for the multiple comparisons between the results obtained at each timepoint. For this comparison, the results of the eight consecutive timepoints were compared (e.g., between 5 and 15 minutes, 15 and 30 minutes); p < 0.006 was considered as significant. For axial signs, a comparison between each hour was added to the analysis; p < 0.005 was considered as significant.

Results.

Switching DBS “off.”

In the 30 patients who completed the study, switching DBS “off” (figure) led to an increase of total UPDRS motor score from 26.5 ± 11.5 to 56 ± 14.7 (p < 0.001). Seventy-five percent of this increase was reached between 15 and 30 minutes (TI 75), and 90% between 90 and 120 minutes (TI 90).

• Download figure
• Open in new tab
• Download powerpoint

Figure. Evolution of Unified Parkinson’s Disease Rating Scale (UPDRS) motor score and subscores when subthalamic nucleus deep brain stimulation is turned “off” (time 0): average and standard errors (30 patients). The two first data points (squares) are preoperative scores “on” and “off” medication. The last data point (open circle) represents the results at 4 hours extrapolated from the data of 11 patients.

Bradykinesia increased from 17.2 ± 6.2 to 31.1 ± 6.6 (p < 0.001): TI 75 was between 15 and 30 minutes and TI 90 was between 60 and 90 minutes. The bradykinesia subscores increased from timepoint to timepoint up to 90 minutes (p < 0.005).

Tremor increased from 1.8 ± 2.3 to 7.2 ± 5.2 (p < 0.001): TI 75 and TI 90 were within 15 minutes. The tremor subscores increased only between baseline and 5 minutes (p < 0.0001).

Rigidity increased from 3.4 ± 2.9 to 10.7 ± 5.1 (p < 0.001): TI 75 was between 30 and 60 minutes and TI 90 between 60 and 90 minutes. The rigidity subscores increased from timepoint to timepoint up to 90 minutes (p < 0.006).

Axial signs increased from 4.2 ± 3.4 to 7.4 ± 4.0 (p < 0.001): TI 75 was between 30 and 60 minutes and TI 90 between 120 and 150 minutes. The differences in axial subscores from timepoint to timepoint were not significant. However, such significance was obtained when differences from hour to hour were tested up to the end of the observation (p < 0.005).

For the 11 patients observed during 4 hours, the results were similar to those observed for the whole group for the 3 first hours. During the fourth hour the UPDRS increased from 54.2 ± 11.8 to 55.8 ± 11.9 (p < 0.05), bradykinesia from 30.3 ± 5.3 to 30.5 ± 5.3, tremor from 6.4 ± 5.2 to 6.8 ± 5.3, rigidity from 9.2 ± 5.1 to 9.5 ± 4.8, and axial signs from 8.5 ± 3.3 to 9.0 ± 3.3: none of these changes was significant. At 3 hours, the 11 patients had reached 90% of the worsening of all their own subscores except axial signs (75% was reached for these signs). Three patients had accepted first to keep the DBS “off” for 6 hours. They did not worsen over the last 2 hours. UPDRS motor score went from 67.7 ± 14.6 at 4 hours to 68.3 ± 14.6 (p = 0.18) at 6 hours. Rigidity scores went from 12.7 ± 2.5 to 13.3 ± 2.1, p = 0.18. Bradykinesia (33.3 ± 5.1/33.3 ± 5.1), tremor (14.7 ± 5.8/14.7 ± 4.9), and axial signs (7.0 ± 5.6/7.0 ± 5.6) remained unchanged.

Discussion.

We found that only 75% of the UPDRS total motor score worsening occurs during the half hour after switching STN DBS “off.” In addition, the motor signs respond differently to the DBS “off”: the change occurring during the first half hour reflects the full worsening of tremor, but only partially the worsening of bradykinesia and rigidity, and barely one of the axial signs. Patients keep on worsening over the 4 hours after switching DBS “off.” When switching DBS back “on,” the parkinsonian signs improved faster than they had worsened, but in a similar succession: 1) tremor improves first, 2) rigidity and bradykinesia second, 3) axial signs last. Few patients reported that they felt the return of full effect of DBS after 24 to 48 hours, suggesting an even longer delayed effect of DBS. This lasting effect of STN DBS on parkinsonian signs suggests that DBS is not only effective through a direct depolarizing of STN neurons, because the refractory period of stimulated neurons is rarely prolonged over seconds. Hours of effects probably imply other mechanisms such as secondary messengers, long-term potentiation, synthesis of proteins, or other delayed phenomena in the basal ganglia circuitry. The differences in temporal sequences of the STN DBS effects on major PD signs further suggest different pathophysiologic mechanisms underlying these signs.15

Tremor was the most rapidly affected sign, with return of the resting tremor within minutes. Such a rapid evolution has also been observed when thalamic Vim or GPi DBS were switched “off.” This almost immediate effect suggests direct neuro-transmission interference, or perhaps depolarization, in the reverberating loops involved in tremor.16 In contrast to the effect on tremor, stopping STN DBS gradually worsened rigidity and bradykinesia over 1.5 to 3 hours. Bradykinesia and rigidity are the most responsive parkinsonian signs to levodopa and are more specifically improved by STN DBS, in comparison to GPi DBS or Vim DBS.4,5,17⇓⇓ Axial signs (particularly gait and posture) were the slowest to worsen after STN DBS was switched “off.” This evolution is similar to the slow improvement of gait and balance that we observe after STN DBS implantation. Such a delayed and long-lasting effect on gait and axial signs may reflect a tonic long-lasting modification of the basal ganglia activity by STN DBS, probably through connections between the STN and other brainstem nuclei such as the pedunculopontine nucleus and other pontine and medullar reticular formation nuclei involved in gait and balance.8,18⇓

The lasting effect of DBS, whatever the mechanism might be, must be taken into account in the follow-up evaluation of the patients. Currently, the length of observation “DBS off” is not standardized in the literature (1 hour, 2 hours, 12 hours, “overnight”), not even being mentioned by some authors.2,6,19-21⇓⇓⇓⇓ The variability between studies may prevent comparison: the longer the “off” period, the greater the observed DBS effect, with a nonlinear correlation between the two. This situation is comparable to what has been observed with patients off levodopa, with persistence of a long-duration response.22-25⇓⇓⇓ Because of the risks and inconvenience of longer drug holidays, a practically off medication state had to be defined for the purpose of studies.10,11⇓ In our study, patients reached on average 75% of the worsening within 30 minutes after turning DBS “off,” except for axial signs. However, significant worsening of bradykinesia subscores was observed over 2 hours, and total motor UPDRS score increases over 4 hours, reflecting the worsening of axial signs subscores. To assess all parkinsonian signs, an observation of at least 3 hours seems to be required. Indeed, axial signs subscores worsening remains significant 3 hours after switching DBS “off,” although not clearly significant in the fourth hour. We would therefore propose a DBS practically stimulation off score being defined as the UPDRS total motor score observed along 3 hours and reflecting submaximal worsening of all parkinsonian signs in most patients. In our study, only two patients did not stand this length of study, demanding to switch DBS “on” again after 2 hours. The slow progression of UPDRS scores after these 3 hours (the change was barely observable in the patients we examined over 6 hours) does not support longer observations, which may not be bearable for the patients.