Effects of Subthalamic Nucleus Deep Brain Stimulation on Depression in Patients with Parkinson’s Disease

Objective: In this study, we aimed to investigate the effects of STN-DBS on PD patients with different levels of depression and to identify predictors of the effects of STN-DBS on PD depression. Methods: We retrospectively collected data for 118 patients with PD depression who underwent STN-DBS at Beijing Tiantan Hospital. Neuropsychological, motor, and quality of life assessments were applied preoperatively and postoperatively. All patients were divided into two groups according to their HAM-D24 total scores (group I: mild depression; group Ⅱ: moderate depression). A mixed repeated-measure analysis of variance (ANOVA) was performed to investigate whether there were differences in depression scores before and after STN-DBS between the two groups. The changes in depression scores were also compared between groups using ANCOVA, adjusting for gender and preoperative HAMA scores. Logistic regression was performed to identify predictors of STN-DBS’s effects on PD depression. Results: Both groups showed significant improvement in depression symptoms after STN-DBS. Compared with patients in group I, patients in group Ⅱ showed greater reductions in their HAM-D24 total scores (p = 0.002) and in HAM-D24 subitems including cognitive disturbances (p = 0.026) and hopelessness symptoms (p = 0.018). Logistic regression indicated that gender (female) (p = 0.014) and preoperative moderate depression (p < 0.001) patients had greater improvements in depression after STN-DBS. Conclusions: Patients with moderate depression showed better improvement than patients with mild depression. Gender (female) and preoperative HAMA scores are predictors of STN-DBS’s effects on PD depression.


Introduction
Parkinson's disease (PD) is a neurodegenerative disorder characterized by motor symptoms (tremor, rigidity, and slowness of movements) and several non-motor symptoms [1][2][3][4]. Depression, one of the most common non-motor symptoms, is a chronic and commonly reported neuropsychiatric disorder that occurs in 40%-50% of patients with PD [5,6]. PD depression is mainly described as feelings of worthlessness or guilt, a depressed mood, reversible dementia, loss of interest in daily activities, recurrent suicidal tendencies [7][8][9][10], and somatic complaints (muscle tension or sexual dysfunction) [11]. Symptoms of depression can occur at the initial stage of PD motor symptoms and can worsen over time [12], thus negatively affecting quality of life.
To date, several new techniques, for example, deep brain stimulation (DBS), transcranial magnetic stimulation (TMS), and transcranial direct current stimulation (tDCS), are being used for the treatment of PD. DBS is a widely used surgical technique that involves the implantation of multiple-contact electrodes in specific brain regions. Targets of DBS for the treatment of PD include the subthalamic nucleus (STN), external globus pallidus (GPi), pedunculopontine nucleus (PPN), and the substantia nigra pars reticulata (SNr) [13].
Subthalamic nucleus deep brain stimulation (STN-DBS) is a well-established surgical treatment for PD. Although STN-DBS is widely acknowledged to control motor symptoms, including tremor, rigidity, and bradykinesia, the efficacy of STN-DBS in treating PD depression remains controversial. Reports have shown that STN-DBS significantly ameliorates symptoms of depression post-operatively, and its efficacy is greater 6 months after the operation [14,15]. Some articles have concluded that STN-DBS shows little evidence of ameliorating depression, whereas Follett et al. have reported the worsening of depressive disorder after STN-DBS [16]. In another article, suicidal behavior was found to occur within 3 years postoperatively, and the suicide rate was elevated in patients with severe depression [17], thus suggesting that patients with severe depression are not candidates for surgery. Furthermore, the authors of one study declared that PD patients become angry more easily after STN-DBS [18].
Several factors may explain the reported differences in the efficacy of STN-DBS for PD depression. First, the participants in the experiments might have had different baseline data (varying degrees of depression). Moreover, differences in study design, assessment conditions (on medication vs. off medication), and different stimulator settings might have led to contradictory results. However, despite these possibilities, the factors associated with changes in postoperative depression remain unclear. Therefore, our study aimed at verifying the exact effects of STN-DBS on PD depression by dividing patients with PD into groups according to the severity of their depression and further investigating possible predictive factors of STN-DBS's efficacy in treating PD depression.

Patient Selection
We retrospectively collected data from patients at Beijing Tiantan Hospital (Beijing, China) from January 2016 to August 2021. The inclusion criteria were as follows: (1) diagnosis of idiopathic PD according to the United Kingdom Brain Bank criteria [1]; (2) successful performance of bilateral STN-DBS; (3) performance of motor and psychological assessments (preoperative and postoperative) with available postoperative follow-up ranging from 6 to 12 months; (4) Mini-Mental State Examination (MMSE) score > 24; and (5) all patients suffered from depressive symptoms with HAM-D 24 scores > 0. The exclusion criteria were as follows: (1) postoperative complications, such as cerebral hemorrhage and hemiplegia; (2) clear brain structural abnormalities, intracranial tumors, or prior brain surgery other than STN-DBS; (3) revised or replaced DBS leads; and (4) incomplete assessment data; and (5) patients with a total HAM-D 24 score ≥ 35, since major depression is considered a contraindication for STN-DBS. This study was conducted under the approval of the IRB of Beijing Tiantan Hospital. All patients provided written informed consent for their participation.

DBS Electrode Implantation
The standard surgical procedure was as previously described [19,20]. Briefly, quadripolar DBS electrodes (model 3389, Medtronic, Minneapolis, MN, USA, or model L301, Pins Medical, Beijing, China) were implanted under the guidance of a Leksell microstereotactic system (Elekta Instrument AB, Stockholm, Sweden) with pre-surgical MRI under local anesthesia. Intra-operative single-unit recordings and high-frequency stimulation testing were performed to evaluate the optimal locations for permanent electrode implantation. The STN target coordinates for the lower contact were 2-3 mm posterior to the MCP, 12-14 mm lateral to the AC-PC, and 4-6 mm below the inter-commissural line. The electrodes were then connected to an implantable pulse generator implanted in the subclavicular area in patients under general anesthesia. We performed post-operative CT to exclude intracranial hemorrhage, then merged the images with the preoperative MR images to verify the exact locations of the electrodes.

Clinical Assessment
The age, sex, time of onset, duration of disease, and Hoehn and Yahr scale results were recorded for all patients in detail. The levodopa-equivalent daily dose (LEDD) for all patients is presented to provide a better understanding of the dosage of anti-Parkinson's disease drugs. The proportions of PD patients taking antidepressants preoperatively and postoperatively were also recorded. The Unified Parkinson's Disease Rating Scale part III (UPDRS-III) or Movement Disorder Society Unified Parkinson's Disease Rating Scale part III (MDS-UPDRS part III) was applied to all patients to evaluate their motor symptoms. Other related symptom scales were used, including the Hamilton Rating Scale for Anxiety (HAMA), Parkinson's Disease Questionnaire (PDQ-39), and MMSE. PDQ-39 is the most thoroughly validated and extensively used scale assessing the quality of life among patients with PD [21,22], and the MMSE scale is used to assess cognitive function. Motor assessments were applied in two conditions (off medication and on medication) 1 week before surgery preoperatively, and two conditions (off medication/on stimulation and on medication/on stimulation) postoperatively. MDS-UPDRS-III improvement was calculated as (pre-MDS-UPDRS-III [med-off]-post-MDS-UPDRS-III [stim on/med off])/pre-MDS-UPDRS-III [med-off] ×100%. We selected the assessment results of the last return to the hospital for follow-up analysis (at least 6 months postoperatively). All assessments were conducted by a movement disorder specialist in our center, and all motor evaluations were video-recorded. Off-medication assessment was performed at least 12 h after withdrawal from dopaminergic medications, and on-medication assessment was performed 1 h after medication administration. Post-operative CT was performed at 1 month to verify the exact locations of the electrodes by merging the images with the preoperative MR images and was used to guide electrode choice during the adjustment of stimulation parameters. At each assessment time point, patients were required to visit the outpatient programming service to ensure that all assessments were conducted under optimal stimulation parameters.

Statistical Analysis
All continuous variables are presented as mean ± SD, and categorical variables are presented as percentages. Kolmogorov-Smirnov tests were conducted to determine whether continuous variables were normally distributed. Differences in demographic characteristics were compared between group I (mild depression) and group II (moderate depression) using independent t-tests, nonparametric tests, and chi-squared tests where appropriate.
To investigate whether there were differences in the severity of depression after STN-DBS between the two groups, a mixed repeated-measure analysis of variance (ANOVA) was performed, with group (group I and group II) as the between-group factor and time (before and after DBS) as the within-group factor, followed by Bonferroni's post hoc tests.
Then, the change in the severity of depression (HAM-D 24 total scores and subitems) between the preoperative and postoperative periods was calculated for each participant and compared between groups. Comparisons were analyzed first without covariates, using an independent t-test. Second, an analysis of covariance (ANCOVA) was conducted to detect differences in the change in depression scores between the preoperative and postoperative periods, while adjusting for covariates. Covariates of gender and preoperative HAMA scores were selected due to their association with depression scores. Apart from the change in depression scores, comparisons were also made with the changes in motor symptoms (MDS-UPDRS-III), cognition (MMSE), and quality of life (PDQ-39).
Finally, for differences between groups, categorical variables were compared using the chi-squared test. We applied independent-samples t-tests for between-group comparisons when data were normally distributed and nonparametric tests (Mann-Whitney U test) when the data had a skewed distribution. Independent predictors of depression amelioration were analyzed with logistic regression, and family-wise error comparison was also applied for multiple comparisons. Two-tailed p-values below 0.05 were considered to indicate significant results. All statistical analyses were conducted in SPSS 24 (IBM, Chicago, IL, USA).

Patients and Baseline Characteristics
Data for 146 patients with PD were briefly reviewed, and 22 patients were excluded on the basis of the inclusion criteria. Among the 124 patients who met the criteria, two patients were lost to the 12-month follow-up postoperatively, one patient had missing or incomplete psychological evaluation data, and three patients had an accident (trauma or cardiovascular or cerebrovascular disease). The 118 patients who met the criteria were included. The shortest postoperative evaluation time point was 6 months and the longest was 12 months, with an average of 9.78 months in group I (9.78 ± 1.77 months) and 9.88 months in group II (9.88 ± 1.68 months). No significant difference was found in the postoperative follow-up between the two groups. For the evaluation of motor symptoms, 42 of 118 patients were evaluated with the UPDRS-III scale, whereas the rest were evaluated with the MDS-UPDRS scale. We conducted score conversion as described previously [32] to transform the UPDRS-III scale into the MDS-UPDRS scale. According to the HAM-D 24 scale, 78 of the 118 patients were assigned to group I, and 40 patients were assigned to group II. For baseline data in the two groups, female patients had more severe depression symptoms than male patients. Patients in group I had lower HAM-D 24 and HAMA scores than patients in group II, and the PDQ-39 scores were also lower in group I than in group II, thus indicating that patients in group II had more severe depression and lower quality of daily activities. In group I, nine patients (11.54%) were taking antidepressants preoperatively and four patients (6.41%) postoperatively, whereas in group II, 10 patients were taking antidepressants preoperatively and six patients postoperatively. No significant differences were found preoperatively (p = 0.06) or postoperatively (p = 0.14) between the two groups. No other differences were found between groups. Baseline data are shown in Table 1.

Effects of STN-DBS on PD Depression, Motor Symptoms, Cognition, and Quality of Life
The results of the two-way repeated-measures ANOVA are presented in Table 2. The ANOVA revealed a significant main effect of time (F = 66.34, p < 0.001), with higher HAM-D 24 total scores in the preoperative period (18.42 ± 6.51) than in the postoperative period (13.18 ± 8.01), and a main effect of group, with group I being lower (12.71 ± 5.25) than group II (20.28 ± 8.77) (F = 118.6, p < 0.001). The interaction effect demonstrated that PD patients in group II had greater declines in their HAM-D 24 total scores (F= 12.46, p < 0.001) and subitems anxiety/somatization (F = 5.01, p = 0.027), cognitive disturbances (F = 10.63, p = 0.001) and hopelessness symptoms (F = 4.63, p= 0.033). Post hoc analyses indicated that the postoperative HAM-D 24 total scores were significantly decreased compared to the preoperative period in both group I (p < 0.001) and group II (p < 0.001). For HAM-D 24 subitems, group I showed significant postoperative declines compared to the preoperative period in anxiety/somatization (p = 0.020), circadian fluctuations (p = 0.002), retardation symptoms (p = 0.012), sleep disturbances (p = 0.030), and hopelessness symptoms (p = 0.016), whereas group II showed significant postoperative declines compared to the preoperative period in anxiety/somatization (p < 0.001), cognitive disturbances (p < 0.001), circadian fluctuations (p = 0.026), retardation symptoms (p < 0.001), sleep disturbances (p < 0.001), and hopelessness symptoms (p < 0.001). The change between the preoperative and postoperative periods in terms of depression (HAM-D 24 total score and its subitem scores), as well as motor skills (MDS-UPDRS-III score), cognition (MMSE score), and quality of life (PDQ-39 score) are demonstrated in Table 3 1-7). Thus, the result of the chi-squared tests showed that the proportion of patients with decreased depression levels after STN-DBS was higher in group II than in group I (p = 0.002). Independent t-tests without covariates indicated that PD patients in group II showed greater reductions than patients in group I in terms of the HAM-D 24 total score (p = 0.003) and HAM-D 24 subitems, including anxiety/somatization (p = 0.027), cognitive disturbances (p = 0.010), and dopelessness symptoms (p = 0.033). Furthermore, adjusting for gender and preoperative HAMA scores resulted in fewer differences between group I and group II, with group II exhibiting a greater reduction in the HAM-D 24 total score (p = 0.002) and HAM-D 24 subitems, including cognitive disturbances (p = 0.026) and hopelessness symptoms (p = 0.018). In addition, the changes between preoperative and postoperative periods in motor symptoms (MDS-UPDRS-III), cognition (MMSE), and quality of life (PDQ-39) were also compared between groups; however, no significant differences were found in independent t-tests without covariates or ANCOVA adjusting for gender and preoperative HAMA scores. We also calculated the Relative Change as well as Effective Size in both groups. Relative Change= (mean follow-up -mean baseline)/mean baseline. Effective Size= (mean baseline -mean follow-up)/SD change score. The improvement of Anxiety/Somatization, Cognitive Disturbances, Retardation Symptoms, Sleep Disturbances and Hopelessness Symptoms in Group II were classified as medium while others as small (Table 4).

Preoperative Predictors of the Effects of STN-DBS on PD Depression
To predict the efficacy of STN-DBS in patients with Parkinson's disease with different levels of depression in clinical practice, postoperative predictors of Parkinson's disease depression are necessary. We chose age; sex; age of onset; disease duration; and preoperative factors, including the LEDD, MDS-UPDRS score (med off), HAMA score, HAM-D 24 score, and PDQ-39 scale score, as possible predictors. After correction, sex (0.35 (0.15, 0.81), p = 0.014) and the preoperative HAM-D 24 score (4.92 (2.00, 12.13), p < 0.001) were significant in both univariate and multivariate analyses, thus suggesting that sex and preoperative depression were possible predictors of postoperative depression after STN-DBS (Table 5). PD patients with more severe preoperative depression levels were more likely to gain better effects postoperatively. Female PD patients exhibited better improvements in their depression scores after STN-DBS (0.35 (0.15, 0.81), p = 0.014). In contrast, preoperative LEDD and PDQ-39 scores were significant according to the univariate analysis but not according to the multivariate analysis. Other factors were not significant according to both the univariate analysis and multivariate analysis. All data were presented as mean ± SD. p-values < 0.05 were considered to be significant and are marked with * Two-way repeated-measures ANOVA, with group (group I and group II) as the between-group factor and time (before and after DBS) as the within-group factor, was applied to analyze the differences in the HAM-D 24 total scores and HAM-D 24 subitems, followed by Bonferroni's post hoc tests. Anxiety/somatization (6 items: psychic anxiety, somatic anxiety, gastrointestinal symptoms, hypochondriasis, insight, and general symptoms); weight loss (1 item); cognitive disturbances (6 items: self-guilt, suicide, agitation, depersonalization and derealization, paranoid, and obsessive-compulsive symptom); circadian fluctuations (1 item); retardation symptoms (4 items: depression, work and interests, retardation, and sexual symptoms); sleep disturbances (  Continuous variables are presented as mean ± SD and categorical variables are presented as percentages. a chi-square test, unindicated comparisons were conducted using independent t-test and ANCOVA adjusting for covariates of gender and preoperative HAMA scores. p-value < 0.05 is considered to be significant and is marked with * Anxiety/somatization (6 items: psychic anxiety, somatic anxiety, gastrointestinal symptoms, hypochondriasis, insight, and general symptoms); Weight loss (1 item

Discussion
In this study, we retrospectively analyzed the effects of STN-DBS on depressive symptoms in PD, as well as the possible preoperative predictors of the effects of STN-DBS on depressive symptoms in PD. STN-DBS significantly decreased HAM-D 24 scores in patients with PD depression by an average of 10 months after surgery. Interestingly, patients with different levels of depression had different postoperative outcomes. The amelioration of depressive symptoms was greater in patients with moderate depression than in those with mild depression. Thirty-one patients out of 78 patients (39.74%) in group I had decreased depression levels, decreasing from mild (HAM-D 24 score 8-20) to non-depression, whereas 28 of 40 patients (70%) in group II moved from the moderate category (HAM-D 24 score 21-34) to the mild (HAM-D 24 score 8-20) or non-depression (HAM-D 24 score 1-7) categories. PD patients with more severe preoperative depression tended to be more sensitive to stimulation caused by STN-DBS than patients with mild depressive symptoms and had more space for improvement. We speculated that differences in sensitivity to stimulation were responsible for the different results between the two groups. Simultaneously, the major subitem contributors differed among patients with different depression levels. Furthermore, regression analysis on possible preoperative predictors of postoperative depression amelioration indicated that preoperative HAM-D 24 scores (depression level) and sex predicted the efficacy of STN-DBS in treating depressive symptoms in PD.
In agreement with our results, several recent studies have demonstrated that STN-DBS shows a trend toward the amelioration of PD depression. For example, Huang and colleagues indicated that after STN-DBS, patients' Beck Depression Inventory scale scores decreased significantly, thus indicating the significant alleviation of depressive symptoms. As previously reported, the generation of the depressive symptoms of PD is associated with the limbic system, the ventromedial prefrontal cortex, and the dorsolateral prefrontal cortex [33,34]. In a resting-state functional MRI (rs-fMRI) study of depression by Wen and colleagues, compared to healthy controls, patients with PD were found to have increased neural activity in prefrontal regions and decreased functional connectivity between prefrontal and limbic structures [35]. Research has demonstrated that STN-DBS influences the connectivity among the motor-associated cortex, thalamus, and cerebellum, thereby suggesting that structures influenced by STN-DBS are partly associated with areas involved in depression [36,37]. In contrast, studies have shown that changes in glucose metabolism induced by STN-DBS affect the levels of ketamine, which has rapid antidepressant and anti-anhedonic effects in brain structures [38,39]. This finding may provide a reasonable explanation for why STN-DBS eases depressive symptoms in PD.
Some previous studies have concluded that, contrary to our findings, STN-DBS worsens depressive symptoms in patients with PD. Therefore, we further investigated the possible reasons for these disparate results. First, the different scales used across studies focus on different sub-symptoms of depression, and the symptoms improved by STN-DBS might differ from the symptoms focused on in other scales that have been used [40]. Research conducted by Strutt and colleagues, using the Beck Depression Inventory scale, which focuses on cognitive-emotional symptoms (self-disappointment and self-criticalness) instead of somatic symptoms of depression, has shown an increase in depression postoperatively [41]. In addition, evidence has already indicated that STN-DBS can enable the dosage of levodopa medication to be decreased [29]. A rapid decrease in levodopa medication has been suggested to cause the deterioration of depressive symptoms [42].
Moreover, studies have concluded that major depressive disorder is characterized by different involvement of the two hemispheres; however, these results require further research [43,44]. In the initial stage of Parkinson's disease, symptoms appeared on the left or right side of the body. Different sides of the initial symptoms of the disease could have resulted from abnormalities in different cerebral hemispheres. A lot of PD patients have more severe symptoms in one limb than the other. This may also relate to different hemispheres. Our study focused on the changes in the whole brain induced by STN-DBS, rather than unilateral hemispheres. Further attention should be paid to the effect of each unilateral cerebral hemisphere on depression. By exploring the effects of STN-DBS in ameliorating depressive symptoms and predicting the efficacy of STN-DBS in treating depression in patients with PD, our study may help clinicians and nurses to arrange clinical treatment methods according to the status of each patient's illness. In addition, this study may provide a reference for decision-making and the development of individualized treatment plans, and may help to alleviate patient stress.
Several limitations of our study should be noted. First, the number of patients enrolled in this experiment was 118; thus, the sample size was relatively limited. More patient information must be collected to improve the data statistics. Simultaneously, most participants in the experiment were followed up for no more than 2 years; thus, the data may not reflect the depression of patients with STN-DBS after longer periods of time and consequently might have resulted in inaccurate conclusions. Therefore, a longer follow-up period is necessary. Second, as described above, some studies have suggested the differing involvement of the two hemispheres in depression, but our study focused on the effects on the whole brain. Thus, further attention should be paid to the differences between hemispheres through separate analysis. Third, different scales were used to assess patients' motor function, namely, UPDRS and MDS-UPDRS, because of differences in the timing of patients' assessments. Although we used a validated formula to convert the total scores between scales, the results of our study might possibly have been influenced by this. Finally, we only recorded whether the patients were taking antidepressant drugs and we were unaware of the types and dosages of antidepressants that patients used. In future studies, we will keep more detailed records of the antidepressant usage of PD patients with depression.

Conclusions
In this study, we concluded that STN-DBS could improve depressive symptoms in patients with Parkinson's disease. We innovatively divided the PD patients into two groups according to their levels of depression (group I and group II) and found that patients with more severe depression preoperatively tended to have better improvements after STN-DBS. Furthermore, we performed logistic analysis and found that preoperative depression and gender were predictive factors of postoperative depression outcomes. Our findings could help clinicians to better manage depressive symptoms in PD patients with depression. Follow-up studies should pay more attention to functional imaging or electrophysiology methods in order discover the specific neural mechanisms underlying these findings.  Institutional Review Board Statement: All procedures performed in studies involving human participants were con-ducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1975 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all the patients included in the study.

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy regulations for patients.