Associations Between Bilateral Subthalamic Nucleus Deep Brain Stimulation (STN-DBS) and Anxiety in Parkinson’s Disease Patients: A Controlled Study

Traditionally, Parkinson’s disease (PD) is considered a movement disorder. However, there is increasing evidence to indicate that PD is a multidimensional disease. In addition to motor deficits, it is also associated with a number of non-motor symptoms, including anxiety, which can precede the classical motor features of PD by years; these contribute substantially to a deterioration in quality of life for PD patients.^1,2 Epidemiological studies have reported that the prevalence of anxiety in PD ranged from 5% to 69%.^3,4 This variability in results may be caused by the use of different assessment tools and methods. The anxiety of the PD patient cannot be detected unless there are severe clinical manifestations.³ It not only worsens the symptoms of PD,⁵ but also makes the PD patient “afraid of stepping forward due to the fear of tumbling.”⁶

Although the relationship between dopaminergic transmission and clinical anxiety disorders is complex and poorly understood,⁷ there is evidence that anxiety in PD is associated with abnormalities in dopaminergic transmission.^8,9 Dopamine (DA)-depleted rodents manifest increased anxiety-like behaviors.^10,11 Furthermore, recent evidence suggests that the noradrenergic and serotonergic systems may play a more significant role in the manifestation of PD-related anxiety than previously thought.¹² According to the Braak et al.¹³ and Del Tredici et al.¹⁴ staging of PD pathology, serotonergic cell loss in the raphe nuclei is evident before nigrostriatal dopaminergic degeneration.¹⁴ Noradrenaline (NA) dysfunction is also likely to occur before pronounced degeneration of DA neurons.¹³ In general, PD-related anxiety seems to be related to the dopaminergic, noradrenergic, and serotonergic systems and is manifested before the motor dysfunction of PD. On the other hand, the theory related to the etiology of anxiety symptoms in PD argues that PD-related anxiety is “reactive” and secondary to the psychosocial stress of a chronic disease and the associated disability, as they are usually diagnosed after the PD.¹⁵

Subthalamic nucleus (STN) deep brain stimulation (DBS) is known to be an effective treatment for motor symptoms of PD. An increasing number of studies are now focusing on the influence of STN-DBS on non-motor symptoms. Most researchers have reported that the STN-DBS worsens cognition,^16–18 apathy,¹⁹ and impulse-control disorder ^16,20–22 in PD patients. Researchers are increasingly bound up in the discussion of the influence of STN-DBS on depression in PD patients;^23–26 however, few researchers have focused specifically on changes in PD-related anxiety after STN-DBS; most only mention incidental changes in anxiety as a part of follow-up research. In a 72-case, 15-month follow-up study by Castelli et al.²⁷ and a 41-case, 12-month follow-up study by Thierry et al.,²⁶ it was found that the anxiety of the STN-DBS PD patients was stable during the pre- and post-operative follow-up period. However, in a 33-case, 15-month follow-up study by Kalteis et al.,²⁸ anxiety was reported to be significantly improved. The common limitation of these studies is the lack of a control group, which consequently meant that the researchers were unable to minimize the influence of other effects.

To overcome this limitation and try to elucidate the influence of STN-DBS on PD-related anxiety, the current study compared the changes in anxiety within a DA (Medication) control group and a bilateral STN-DBS group. Based on the probable pathological mechanism of PD-related anxiety, this article first hypothesized that anxiety might be decreased in STN-DBS patients because of the pronounced improvement of motor functioning associated with STN-DBS. However, because of the unchangeable nigrostriatal dopaminergic degeneration in PD, the second hypothesis is that STN-DBS cannot change PD-related anxiety in the long term. Furthermore, PD-related anxiety might continue to worsen with the progression of PD disease.

Method

Subjects

The experimental subjects were patients who received bilateral STN-DBS implantation between January 2008 and June 2009 in the Research Center of Functional Brain Disease (a research institute of the Department of Neurosurgery, Tangdu Hospital, Fourth Military Medical University) and who agreed to participate in the study. The inclusion criteria were a diagnosis of idiopathic PD (U.K. brain bank); Hoehn-Yahr (modified) Grades ≥2.5; normal preoperative brain MRI; and preoperative improvement rate of the Madopar Impact Test ≥40%. The exclusion criteria were: Mini-Mental State Exam (MMSE) score ≤24; any distinct mental disorders detected by psychological assessment and psychiatrists; any anxiolytic drugs being used; and any normal surgical contraindications. During the study period, the following subjects were rejected: those whose combination of antiparkinsonian medications exceeded about 10 mg/day of levodopa equivalent; those who needed the treatment of antipsychotics, anxiolytics, and antidepressants; those whose contact positions of the STN-DBS needed to be readjusted for inaccurate location or drifting; those who needed removal of the STN-DBS for any reason; and, for members of the Medication control group, those who wanted to accept the STN-DBS implantation.

The study applied the paired-control method. Once a PD patient was enrolled as a member of the STN-DBS group, another PD patient from the outpatient databank of our center, who was receiving pharmacotherapy only, was paired to an STN-DBS patient and assigned to the Medication group. The pairing criteria were matching gender and Hoehn-Yahr (modified) grade, age (within 1 year), medical history (within 1 year), years of education (within 1 year), and the daily levodopa equivalent of antiparkinsonian medications (plus-or-minus 100 mg/day).

All patients were told that the levodopa equivalent of their current antiparkinsonian medications would not change during the study period. As this study aimed to investigate the optimal therapeutic doses of levodopa, and as a reward for participation in the study, levodopa was provided to subjects free of charge. Patients were informed clearly that they had the right to withdraw from the study at any time.

The patients were unaware of the aims of the study, and the assessors and program-controllers of the DBS were blind as to patient-group assignment. Furthermore, the designer of the study did not participate in the implementation or assessment procedures. This study was approved by the ethics committee of the Fourth Military Medical University, China.

Assessment Scales

The Hamilton Anxiety Scale (Ham-A; 14-item version) and the State–Trait Anxiety Inventory (STAI) were used to assess the severity of anxiety. Three of the questions in the Ham-A relate to the inner experience of anxiety, such as inquietude, tension, and fear; one is about depression experience; the rest relate to clinical manifestations of anxiety, such as agrypnia, memory deterioration, other physical symptoms (including those that relate to musculature, sensory system, cardiovascular system, gastrointestinal tract, respiratory tract, genitourinary system, and vegetative nervous system), and general performance when talking. The higher the Ham-A scores, the worse the anxiety. The STAI is comprised of two subscales: state anxiety (S–AI) and trait anxiety (T–AI); with 20 items each. S–AI is mainly a measure of current anxiety experience, whereas T–AI is mainly an assessment of recent past/general anxiety experiences; the higher the scores of the two subscales, the greater the anxiety symptoms.

The third part, the motor component of the Unified Parkinson's Disease Rating Scale (UPDRS-III), was used to assess the severity of motor symptoms in the PD patients; higher scores are indicative of more severe PD motor symptoms. The PDQ-39 assesses the quality of life for PD patients; a higher score denotes a worse quality of life.

Evaluation Time-Points

The patients were assessed in the “On” state of levodopa. In order to minimize the influence of operation-related stress, the pre-operative assessment was conducted 1 month before the operation. To assess the post-operative edema reaction and DBS power–On stress reactions, the study carried out the first post-operative assessment at the 3rd week after the operation; that is, 1 week before Stimulator Power–On. The second postoperative assessment was carried out at the 5th week after the operation/1 week after the Stimulator power–On, to observe the short-term changes in anxiety after Stimulator power–On. To observe the transitory and long-term changes in anxiety after Stimulator power–On, subsequent assessments were carried out at 2 months post-operatively/1 month after Stimulator power–On, 4 months post-operatively/3 months after Stimulator power–On, 7 months post-operatively/6 months after Stimulator power–On, and 13 months post-operatively/12 months after Stimulator power–On. The evaluation time-points for the Medication group were strictly synchronized with the STN-DBS group. Two psychologists with clinical qualifications approved by the National Ministry of Health and ≥2 years of clinical experience were responsible for the mental and scale assessments. The average of the scores ascertained by the two assessors was taken as the final result for each patient.

Surgical Procedures

All of the PD patients in the STN-DBS group underwent the bilateral STN-DBS operation under the guide of magnetic resonance imaging (MRI). First, the PD patients had a Leksell stereotactic frame installed under local anesthesia. An MRI (GE 3.0 T) with thin slice (2-mm thickness, 0-mm interval) was then carried out. The imaging data were transferred to Elekta SurgiPlan software for target location. In accordance with the Schaltenbrand-Wahren Brain Anatomy Atlas and using the coronal images, the target lateral dorsal part of the STN was directly localized on the axial image, and the coordinate values of the target were calculated. Two burr-holes were drilled in front of the coronal suture and 2.5–3 cm. perpendicular from the midline bilaterally. The intra-operative Medtronic Leadpoint microelectrode recording system was used to record electrophysiological signals of STN neurons to help locate the target accurately. An electrical stimulation test was used routinely to observe efficacy and side effects. Once a satisfactory test result was obtained, the DBS quadripolar electrode (model 3389; Medtronic Inc., Minneapolis, MN, U.S.A.) was implanted and fixed at the burr-hole. After the electrical stimulation test had been repeated to confirm the therapeutic effect and the absence of side effects, an extension wire (Lead 7482-51) and an implantable pulse-generator (Kinetra 7428; Medtronic, Inc.) was implanted in a subcutaneous pocket, under general anesthesia. One month post-operatively, the pulse-generator was turned on and programmed with the following parameter ranges: pulse width at 60 μS, 90 μS, and 120 μS; frequency at 130–185 Hz; and voltage at 1.4–3.5 V.

Statistical Methods

SPSS 11.0 software was used for data processing and statistical analysis. Mean values were used in the calculation of the bilateral STN-DBS stimulating parameters of voltage, pulse-width, and frequency. The data in the study are expressed as mean (standard deviation [SD]). Statistical methods included analysis of variance, paired-samples t-tests, independent-samples t-tests, correlation analysis, and regression analysis. For comparisons between the two groups, p<0.05 was considered significant. For comparisons of multiple time-points, p<0.01 was considered significant.

Results

Demographic Data for the Two Patient Groups

During the study period, information was obtained from a total of 36 pairs of patients. Five pairs failed to complete the 14-month follow-up. In two pairs, the patients in the Medication control group chose STN-DBS implantation treatment. In one pair, the dose of levodopa was significantly adjusted. In the other two pairs, the electrode was removed because of infection. Valid data were obtained for 31 pairs of cases for the final analysis (efficiency: 86.11%). The demographic data for the two groups of patients are shown in Table 1. There were no statistical differences in demographics between the STN-DBS group and the Medication group (p>0.05).

3.2. Stimulation Parameters and Assessment Scale Data for the Two Groups at Various Time-Points

The results of the average-difference tests are shown in Table 2. There were no differences in the averages of the data for the two groups at the baseline level of 1 month pre-operatively and 3 weeks post-operatively.

Comparisons Between the STN-DBS Group and the Medication Group at Various Time-Points

Compared with the Medication group, the UPDRS-III and PDQ-39 of the STN-DBS group showed improvement after 1 week Stimulator power–On (p<0.05); the S–AI improved significantly at 1 week and 1 month after Stimulator power–On (p<0.05), but the difference was not significant at the subsequent time-points (p>0.05). The T–AI and Ham-A scores became worse after 3 months with Stimulator power–On (p<0.05).

Comparisons of the Stimulation Parameters and Assessment Scales at Various Time-Points for the STN-DBS Group

Compared with 1-month pre-operation and the 3rd week post-operatively, the UPDRS-III and PDQ-39 scores for the STN-DBS group were improved at 1 week and at all subsequent time-points after Stimulator power–On (p<0.05). The S–AI scores showed improvement from 1 week to 6 months after Stimulator power–On (p<0.05), but then anxiety worsened at the 12th month after Stimulator power–On (p<0.05). The T–AI and Ham-A showed significant deterioration from the 3rd month to the 12th month after Stimulator power–On (p<0.05).

The voltage and pulse-width of the STN-DBS group increased significantly at 1 month and at all subsequent time-points after Stimulator power–On, as compared with 1 week after Stimulator power–On (p<0.05). The frequency increased significantly from the 3rd month to the 12th month after Stimulator power–On (p<0.05), whereas no significant change was observed at the 1st month (p>0.05).

Comparisons of the Stimulation Parameters and Assessment Scales at Various Time-Points for the Medication Group

Compared with the pre-operation and Stimulator power–On time-points, the UPDRS-III scores for the Medication group were poor at all time-points after the 5th month of the study (p<0.05). The scores for the PDQ-39 were significantly poorer at the 14th month of the study (p<0.05). The S–AI, T–AI, and Ham-A scores for the Medication group showed no significant changes at any time-points during the study (p>0.05).

Correlations Between Stimulation Parameters and Assessment Scales

The results of the correlation coefficients and significance tests between the UPDRS-III, PDQ-39, and stimulation parameters, and the S–AI, T–AI, and Ham-A are shown in Table 3.

For the Medication group, both the UPDRS-III and PDQ-39 were positively related to S–AI, but not to the T–AI or Ham-A during the study period p<0.05).

For the STN-DBS group, the UPDRS-III was positively related to the S–AI at and before the 3rd month after Stimulator power–On (p<0.05). The PDQ-39 was positively related to the S–AI at and before 3 weeks postoperatively (p<0.05), but was uncorrelated at subsequent time-points (p>0.05). The UPDRS-III and PDQ-39 were not related to the T–AI and Ham-A during the study period (p>0.05). The S–AI was not related to voltage, frequency, or pulse-width after Stimulator power–On, whereas both the T–AI and Ham-A were positively related to voltage and pulse-width (p<0.05).

Regression Analysis for the Influence of Stimulation Parameters on Anxiety Scales in the STN-DBS Group

We analyzed and tested the significance of the standard regression equation and standard regression coefficients in accordance with the findings of the correlation analyses. The UPDRS-III and PDQ-39 were the independent variables, and the S–AI was the dependent variable. Voltage and pulse-width were independent variables, and the Ham-A and T–AI were dependent variables. The results are shown in Table 4.

The effects of the UPDRS-III and PDQ-39 on the S–AI binary standard regression equation were established at 1 month pre-operation and the 3rd week post-operatively (1 week before Stimulator power–On). The effect of the UPDRS-III on the S–AI standard unitary regression equation was established from 1 week to 3 months after Stimulator power–On. The effect of voltage and pulse-width on the T–AI binary standard regression equation was established during 12 months after Stimulator power–On. The effects of voltage and pulse-width on the Ham-A binary standard regression equation were established during 12 months after Stimulator power–On.

Discussion

The results of this study revealed that the UPDRS-III scores progressively increased for the Medication group, indicating that the symptoms of PD continued to worsen when using the unchanged dosage of dopamine during the study period. Also, for the STN-DBS group, wherein the parameters of the electrical stimulation were continually adjusted and the dosage of dopamine was unchanged, the symptoms of PD became progressively worse. These findings both indicate that the progression of PD cannot be controlled by dopamine or STN-DBS.

The study found that S–AI was related to motor symptoms and life quality pre-operatively and 4 months post-operatively for the STN-DBS group, but, in the Medication group, this correlation existed during the entire study period. This indicates that state-anxiety was influenced by the severity of the motor symptoms and the level of life quality. The results therefore verify the “reactive and secondary” theory of PD-related anxiety ¹⁵ and, to some extent, support the first hypothesis of this article (i.e., that PD-related anxiety may be decreased in STN-DBS patients because of the improvement of motor functioning associated with STN-DBS surgery). The motor symptoms of PD were fluctuating, and levels of anxiety paralleled these fluctuations. The lack of a correlation between S–AI and motor symptoms after 4 months postoperatively for the STN-DBS group may be because of changes in the STN-DBS stimulating parameters. This study also revealed that state-anxiety was unrelated to the STN-DBS stimulating parameters. One might infer from this that the stimulating parameters were not related to the symptoms of PD.

The more interesting result was that trait-anxiety and Ham-A scores were correlated with the changes in voltage and pulse-width during the study period. This result conflicted with the second hypothesis (i.e., that STN-DBS will not change PD-related anxiety in the long term because of the unchangeable nigrostriatal dopaminergic degeneration in PD). However, when the hypotheses were formulated, we only considered the unceasing nigrostriatal dopaminergic degeneration of PD patients, and neglected the impact of the STN-DBS or the stimulating target (STN).

Recent research has confirmed that within human and primate brains, there are five circuits that connect the cortex to the basal ganglia, the thalamus, and back to the cortex. The associative circuit and limbic circuit are closely related to cognitive, emotional, and behavioral regulation. The STN is, anatomically, at the center of these loops, and, as such, is an effective regulator of these circuits.^29,30 The STN can be anatomically divided into three functional subregions: the dorsolateral motor subregion, the ventromedial associative subregion, and the medial limbic subregion, which are related to motor functions, the associative circuit, and the limbic circuit, respectively.^29,31–34 Neuroanatomical research³⁵ has shown that the middle part of the STN connects to the limbic part of the globus pallidus and striatum and also links to the prefrontal lobe and cingulate gyrus. These connections suggest that the STN is involved in the information-processing of the limbic system. Greenhouse et al.³⁶ showed that stimulation of the dorsal subregion led to a more significant improvement of motor symptoms, as compared with stimulation of the ventral subregion, but it induced negative emotions much more often.

The best improvement in motor symptoms is produced when one stimulates the dorsolateral subregion of the STN in PD patients,^36,37 which was chosen as the target region in the current study. If continuously stimulated, the dorsolateral subregion of the STN may trigger a variety of moods, such as panic and worry, and bring about physical changes symptomatic of anxiety, such as a rapid heartbeat, muscle tension, and shivers. Also, as the STN connects to the limbic system, the stimulated STN will influence the limbic system, which may affect the amygdaloid nucleus, and instigate moods such as anxiety.

The stimulation parameters of DBS include voltage, frequency, pulse-width, and the selection of contacts. In this study, the dorsolateral motor subregion of the STN was used as a stimulating target for all cases, and all stimulating patterns were single-contact stimuli. The position of the contacts was identified by the effect of the motor symptoms via programming-control of the DBS and the post-operative MRI. If the electrodes were located inaccurately or drifted from the target, the patients were rejected from the study. Therefore, all of the targets that are affected by all electrode contacts can be regarded as the same. For this reason, the contact-selection itself had nothing to do with the anxiety state.³⁸ The voltage and pulse-width of DBS mainly affect the current strength. Frequency mainly refers to the timing of stimulations and has little to do with the current strength. Theoretically, when the voltage and pulse-width are higher, the current that is generated by the DBS stimulator is stronger, and, as a result, the functional brain area that is influenced by the current is larger. If the current grew stronger, it may affect more subregions within or outside the STN. In such cases, anxiety symptoms will increase in severity. Also, the stronger current may be identified by the brain as a harmful stimulation, and subsequently arouse anxiety, an instinctive response to danger.

The financial burden of the DBS device and the battery loss of the pulse-generator were also psychologically distressing to PD patients and their families. Higher voltage and pulse-width meant greater loss of pulse-generator battery power, more money, greater psychological distress, and increased anxiety.

The results of this research differed from those of previous studies.^26–28 The main reason for this may be that the previous studies^26–28 had not controlled for the influence of DA on PD-related anxiety. The present study remedied this shortcoming. Moreover, this study addresses the influence of the stimulating parameters of STN-DBS on PD-related anxiety, the discussion of which has not appeared in the published literature to-date. However, in this type of research, the mean values adopted for the bilateral stimulator’s parameters will be problematic when the differences between the bilateral stimulator’s parameters are great, an effect that we will continue to monitor in future studies. Nevertheless, despite the limitations associated with this research, which also include the short time-course (because of ethical considerations), the small number of cases, and the incomplete random sample, the current study provides some representation of the anxiety experienced by the STN-DBS PD patient population.