Article Text

Download PDFPDF

Research paper
Subthalamic stimulation and neuropsychiatric symptoms in Parkinson’s disease: results from a long-term follow-up cohort study
  1. Marie Abbes1,
  2. Eugénie Lhommée1,2,3,
  3. Stéphane Thobois4,5,6,
  4. Hélène Klinger4,5,6,
  5. Emmanuelle Schmitt1,2,3,
  6. Amélie Bichon1,2,3,
  7. Anna Castrioto1,2,3,
  8. Jing Xie7,
  9. Valérie Fraix1,2,3,
  10. Andrea Kistner1,2,3,
  11. Pierre Pélissier1,2,3,
  12. Éric Seigneuret8,
  13. Stéphan Chabardès2,3,8,
  14. Patrick Mertens9,
  15. Emmanuel Broussolle4,5,6,
  16. Elena Moro1,2,3,
  17. Paul Krack1,2,3,10
  1. 1Movement Disorders Unit, Department of Psychiatry Neurology and Neurological Rehabilitation, CHU Grenoble Alpes, Grenoble, France
  2. 2Grenoble Institut des Neurosciences, Université Grenoble Alpes, Grenoble, France
  3. 3Inserm U1216, Grenoble, France
  4. 4Faculté de Médecine et de Maïeutique Lyon Sud Charles Mérieux, Université de Lyon 1, Université de Lyon, Lyon, France
  5. 5Neurologie C, Hospices Civils de Lyon, Hôpital Neurologique, Lyon, France
  6. 6Centre de Neurosciences Cognitives, CNRS, UMR 5229, Bron, France
  7. 7Institut du vieillissement, Hospices Civils de Lyon, Hôpital des Charpennes, Lyon, France
  8. 8Department of Neurosurgery, CHU Grenoble Alpes, Grenoble, France
  9. 9Neurochirurgie A, Hospices Civils de Lyon, Hôpital Neurologique, Lyon, France
  10. 10Department of Clinical Neuroscience, Faculty University of Geneva, Hôpitaux Universitaires de Genève, Geneva, Switzerland
  1. Correspondence to Professor Paul Krack, Neurology Division, Department of Clinical Neurosciences, University Hospital of Geneva, Geneva 1205, Switzerland; Paul.Krack{at}


Background Reports on behavioural outcomes after subthalamic nucleus deep brain stimulation in Parkinson’s disease are controversial and limited to short-term data. Long-term observation in a large cohort allows a better counselling and management.

Methods To determine whether a long-term treatment with subthalamic stimulation induces or reduces impulse control behaviours, neuropsychiatric fluctuations and apathy, 69 patients treated with subthalamic stimulation are prospectively and retrospectively assessed using Ardouin Scale of Behavior in Parkinson’s Disease before and after 3–10 years of stimulation.

Results At a mean follow-up of 6 years, all impulse control disorders and dopaminergic addiction were significantly decreased, apart from eating behaviour and hypersexuality. Neuropsychiatric fluctuations also significantly improved (ON euphoria: 38% of the patients before surgery and 1% after surgery, P<0.01; OFF dysphoria: 39% of the patients before surgery and 10% after surgery, P<0.01). However, apathy increased (25% of the patients after surgery and 3% before, P<0.01). With the retrospective analysis, several transient episodes of depression, apathy, anxiety and impulse control disorders occurred.

Conclusions Bilateral subthalamic nucleus stimulation was overall very effective in improving impulse control disorders and neuropsychiatric fluctuations in parkinsonian patients in the long term despite a counteracting frequent apathy. Transient episodes of impulse control disorders still occurred within the follow-up. These findings recommend a close follow-up in parkinsonian patients presenting with neuropsychiatric symptoms before deep brain stimulation surgery.

Clinical trial registration NCT01705418;Post-results.

  • apathy
  • dopamine
  • impulse control disorder
  • Parkinson’s disease
  • subthalamic nucleus
View Full Text

Statistics from


Parkinson’s disease (PD) is a neurodegenerative disease rich in motor and neuropsychiatric symptoms.1 It is now known that patients with PD can develop various neuropsychiatric symptoms as early as during the premotor phase2 until the late stages of the disease, which have an important impact on patients3 4 and caregivers.5 According to the level of dopaminergic stimulation, these symptoms can be divided into two main syndromes: the ‘hypodopaminergic syndrome’ which includes apathy, depression and anxiety, and the ‘hyperdopaminergic syndrome’ which includes various behavioural addictions to computer and gardening, creativity, impulse control disorders (‘ICDs’, including pathological gambling, compulsive shopping, excessive eating behaviour and hypersexuality), punding and addiction to dopaminergic drugs, known also under the term of hedonistic homeostatic dysregulation syndrome.6–8

Subthalamic nucleus deep brain stimulation (STN-DBS) is a recognised effective treatment of motor signs and motor complications of levodopa (L-Dopa) therapy in PD.9 10 However, DBS effects on behavioural symptoms and neuropsychiatric fluctuations are still controversial.11 Therefore, deciding for surgery in patients with PD with neuropsychiatric symptoms remains challenging.12 Most concerns are related to several psychobehavioural symptoms that have been reported after STN-DBS. Between hypodopaminergic symptoms, apathy is one of the most frequent complications after STN-DBS both in the early postoperative and long-term follow-up.13 Underlying mechanisms of postoperative apathy seem to imply complex interactions between dopaminergic mesolimbic denervation, dopaminergic treatment and direct effect of DBS on the STN.11 14 15 Moreover, apathy can herald a worsening of cognitive functions belonging to PD progression.16 Concerning the occurrence of depression after STN-DBS, there is no agreement about improvement or worsening after surgery. Indeed, depressive episodes, both transient and chronic, tend to occur more frequently under STN-DBS treatment, but by contrast, the severity of depressive symptoms assessed by mood scales is significantly reduced under STN-DBS compared with best medical treatment.11 In the same way, increased risk for suicidality has been reported in retrospective studies.17 However, recent randomised controlled trials have shown no specific influence of STN-DBS on depression or suicide.10 18 Regarding anxiety, most literature shows either no effect of STN-DBS11 or improvement of the symptoms.19 Between hyperdopaminergic symptoms, existing studies focus on the effect of STN-DBS on ICDs. Since STN-DBS permits to considerably reduce dopaminergic treatment, an alleviation of ICDs and behavioural addictions is expected.12 Indeed, behavioural addictions have been linked to a dopaminergic overdose, especially dopaminergic agonists with selective affinity for D2-D3 dopamine receptors for ICDs.6 Recent prospective studies have shown either no increase or an alleviation of ICDs in STN-stimulated patients,20–22 except for eating behaviour. However, a worsening of pre-existing behavioural disorders, or onset of new behavioural disorders has also been described.23 24 To note, no randomised controlled trial has specifically addressed this issue,25 and available prospective studies report outcomes of neuropsychiatric symptoms in patients with PD with STN-DBS up to 3 years postoperatively.10 25 26 To fill this gap, we have studied the long-term prevalence of hypodopaminergic and hyperdopaminergic symptoms as well as neuropsychiatric fluctuations in a cohort of patients with PD with bilateral STN-DBS followed up to 10 years. We have analysed both the mean outcomes of the whole cohort and the retrospective individual outcomes in a subgroup. In this way, we have been able to pin down single specific aspects and capture transient symptoms, since hypodopaminergic and hyperdopaminergic symptoms are potentially reversible following dopaminergic treatment adaptation.

Materials and methods


Participation was proposed to the whole cohort of 102 patients with PD (n=66 in Grenoble and n=36 in Lyon) that participated to a previous study focusing on the impact of apathy up to 1 year after STN-DBS (the ‘Apathy Study’). Inclusion criteria of the initial cohort have been already published.14

Design of the present study

Clinical assessments were planned before surgery,14 and in a period between 3 and 10 years after surgery (endpoint visit, figure 1), under chronic treatment of medication and stimulation. While drug management was well defined during the first postoperative year,14 medication and stimulation parameters were chosen freely by the usual treating neurologist of the patient (in town, in the local hospital, in the surgical centre, and so on) during the subsequent follow-up. In case of postoperative dementia diagnosed with Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV) criteria, psychobehavioural evaluations were performed with the help of patient’s close relatives. Some clinical assessments were retrospective (see below).

Figure 1

Periods definition for prospective and retrospective assessments. STN-DBS, subthalamic nucleus deep brain stimulation.

All the patients gave written informed consent. A part of this work (baseline evaluation) was supported by ‘Programme Hospitalier de Recherche Clinique Interrégional’ and Servier Pharmaceutical Company. The study is registered with (number NCT01705418).

Prospective assessments

Cognitive function was assessed using the Mattis Dementia Rating Scale. Frontal subcortical deterioration was evaluated using the Frontal Score, composed of verbal fluency tests, graphic and gestual series and the Wisconsin card sorting test.

The Mini International Neuropsychiatric Interview (MINI)27 was used to screen previous and current psychiatric events based on the DSM-IV. In the MINI, ‘Lifetime’ evaluation covers the period comprised between the beginning of adulthood and the day of the assessment and ‘current life’ evaluation concerns the past month. Self-assessments of depression, anxiety and apathy were scored using the Beck Depression Inventory, the Beck Anxiety Inventory and the Starkstein Apathy Scale. Mood and behaviour were assessed using the Ardouin Scale of Behavior in Parkinson’s Disease (Ardouin Scale), comprising 21 items ranging from 0 (no change) to 4 (severe change in behaviour), and grouped in three subscales: hypodopaminergic behaviours (five items: depressed mood, anxiety, irritability, hyperemotivity and apathy); neuropsychiatric fluctuations (two items: non-motor ON euphoria with eventual urge to talk, psychological strengthened dynamism, and so on; non-motor OFF dysphoria with distress, panic, and so on); and hyperdopaminergic behaviours (14 items: hypomanic mood, psychotic symptoms, nocturnal hyperactivity, diurnal somnolence, eating behaviour, creativity, hobbyism, punding, risk-taking behaviour, compulsive shopping, pathological gambling, hypersexuality, dopaminergic addiction and excess in motivation).28 Items with a score equal or superior to 2 were considered abnormal. The time reference of the evaluation is the past month (figure 1).

The summary index of the Parkinson’s Disease Questionnaire (PDQ-39) of quality of life was used to assess the functional impact of parkinsonian motor and non-motor signs.

The Unified Parkinson’s Disease Rating Scale (UPDRS) parts III and IV were used to assess the motor signs and dyskinesia. Before surgery, the OFF condition was assessed after one night medication withdrawal and the ON condition was scored after a suprathreshold dose of L-Dopa.14 Chronic medication (L-Dopa and dopamine agonists expressed in levodopa equivalent daily dose), psychotropic medication (number of patients with an antidepressant treatment, number of patients with a neuroleptic treatment) and parameters of subthalamic stimulation were also recorded.

Retrospective assessments

For the patients followed in Grenoble who could be included in this study (n=48/66), all major psychobehavioural events that occurred from the first year after surgery (follow-up period, figure 1) were recorded retrospectively using the Ardouin Scale. Clinical characteristics, including phenomenology, and number and duration of each episode were collected. We focused on four profiles of patients: (1) patients who never displayed any significant psychobehavioural symptom: ‘normal behavior’; (2) patients who displayed only hypodopaminergic symptoms (once or more times): ‘hypodopaminergic behavior’; (3) patients who displayed only hyperdopaminergic symptoms (once or more times): ‘hyperdopaminergic behavior’; and (4) patients presenting with both hypodopaminergic and hyperdopaminergic behaviours: ‘oscillating behavior’. Suicide attempts and their circumstances were also reported.

Statistical analysis

All prospective data were compared with preoperative measures (test-retest). An exact McNemar’s test was used to compare paired qualitative parameters. For continuous data, a Wilcoxon test was used for paired comparisons. Data were summarised in terms of size and frequency for categorical data, and by mean scores±SD for quantitative data. P values <0.05 were considered statistically significant. Missing data are specified in each result table. Analyses were performed with Statistica (V.7.0). Results of retrospective assessments have only been described.


Among the 102 patients enrolled in the initial study, 33 were lost to follow-up, or refused to participate in the study or died (figure 2). The main clinical characteristics of the study population (n=69) are described in table 1. Median follow-up after surgery was 6 years.

Table 1

Clinical characteristics of the patients at baseline and follow-up

Figure 2

Flow chart of patient’s enrolment. PD, Parkinson’s disease.

Prospective assessments at endpoint visit


Overall efficiency and executive functioning decreased significantly compared with before surgery (table 1). Dementia was diagnosed postoperatively in 7/69 patients (10%).

Psychiatric symptomatology: MINI evaluation

The ‘lifetime’ prevalence of panic disorders decreased significantly after surgery. The ‘current’ suicide risk increased at endpoint visit. No other psychiatric episode significantly differed in frequency between the two evaluations (table 2).

Table 2

The MINI prevalence in percentage of psychiatric disorders concerning both current and lifetime periods at baseline and endpoint visits in 69 patients

Mood and behaviour: Ardouin Scale and self-assessments

At the last follow-up, there was a significant reduction of both neuropsychiatric fluctuations in OFF and ON (table 3).

Table 3

Number of patients displaying a disorder using the Ardouin Scale (score ≥2) before surgery (baseline) and at endpoint visit (3–10 years after surgery)

Overall, hyperdopaminergic behaviours were markedly reduced in the follow-up. Notably hypomania, nocturnal hyperactivity, creativity, hobbyism, pathological gambling, compulsive shopping, risk-seeking behaviour and dopaminergic addiction were significantly less frequent after surgery (table 3). Nevertheless, 12 patients were still affected by excessive eating behaviour, and seven patients by diurnal somnolence. Hypersexuality was observed in three patients at the endpoint visit (vs two patients preoperatively). Psychotic symptoms tended to increase (none before surgery vs four after surgery).

Frequency of depressed mood increased without reaching significance (four patients before vs nine after surgery). During the postoperative period, four patients attempted suicide (5.8%). One patient attempted suicide in the immediate postoperative period (within the 3 months after surgery), whereas the other three patients few years after surgery. No completed suicide was observed. Finally, a significant increase in apathy was observed (2 patients before vs 17 after surgery, ie, 24.6% of the patients) (table 3). This prevalence is quite identical (26.2% of the patients) when considering only the 42 patients displaying a preoperative hyperdopaminergic behaviour (defined by one or several hyperdopaminergic items >1 in the Ardouin Scale).

Self-assessments of depression, anxiety and apathy are summarised in table 1. Apathy and depressive symptomatology increased significantly, whereas anxious symptoms remained stable between the two evaluations.

Medication, motor state and quality of life

When compared with the preoperative assessment, dopaminergic medication was significantly reduced by 55.3%. The UPDRS III scores under chronic treatment condition were significantly higher when compared with the preoperative ON-drug condition (table 1). Concerning psychotropic medication, one patient took an antidepressant at baseline compared with 23 patients at endpoint visit; none of the patients took a neuroleptic molecule before surgery compared with five patients at endpoint. The PDQ-39 Summary Index did not change at endpoint visit compared with baseline.

Retrospective assessments: behavioural events during follow-up period

After surgery, apathy occurred ex nihilo in 22/48 patients, meaning that 45.8% of the patients displayed transiently or more permanently this symptom during the whole 6-year follow-up period. In most cases (15/22), apathy was not associated with depression or anxiety. Six out of the seven patients who developed dementia were apathetic.

After surgery, depression occurred in 13/48 patients (10 ex nihilo, three with preoperative depressions), meaning that a quarter of the patients displayed transiently or more permanently this symptom during the whole 6-year follow-up period. In half of the cases (7/13), depression was associated with apathy. In three other cases, depression was interpreted as reactive to ICD.

Eating behaviour disorder was present in 13/48 patients (27.1%, 9/13 ex nihilo, 4/13 with preoperative eating behaviour disorder). Hypersexuality, compulsive shopping and/or pathological gambling occurred ex nihilo in 14/48 patients (29.2%), and the relapse of these disorders after surgery concerned four patients. Among ICD, hypersexuality was most frequently observed, displayed in 8/48 (16.7%) patients.

Four behavioural profiles were established in the 48 patients. The ‘hypodopaminergic’ profile was encountered in 13 patients before surgery and in 20 patients at endpoint visit. The ‘hyperdopaminergic’ profile was fulfilled by 27 patients before surgery and by 14 patients at endpoint visit. The ‘normal behavior’ profile was observed in eight patients before surgery and 14 patients at endpoint visit.

If we consider only the follow-up period (between the first postoperative year and the endpoint visit), 10 patients presented the ‘hypodopaminergic’ profile, 11 patients presented the ‘hyperdopaminergic’ profile and 9 patients an ‘oscillating’ profile. Finally, 18 patients presented the ‘normal behavior’ profile during the follow-up period.

The mean delay between the baseline and the time of first disorder detection was about 3.5 years for the hypodopaminergic behaviours and about 2.5 for hyperdopaminergic behaviours.

Case reports

Case 1

A man/woman, in his/her fifth decade, diagnosed with PD since 7 years, underwent STN-DBS surgery for invalidating motor fluctuations. L-Dopa equivalent dose was 1200 mg/day, agonist L-Dopa equivalent dose was 300 mg/day. At this time, he/she presented severe dopamine replacement therapy side effects on his/her behaviour that persisted despite repeated medication adaptation attempts: pathological shopping, nocturnal hyperactivity and risk-taking behaviour. One year after surgery, the motor effect of STN-DBS was satisfying, allowing for a reduction of L-Dopa therapy to 400 mg/day and dopamine agonist arrest. Pathological behaviours had disappeared. No apathy or depression was noticed. Five years after surgery, this was still the case and he/she had a good quality of life: he/she was active, well balanced in his/her personal and social life, complaining about a lowering of his/her voice but without any motor complaint. This case illustrates a beneficial effect of STN-DBS on behaviour in the long term.

Case 2

A man/woman in his/her sixth decade, diagnosed with PD since 10 years, underwent STN-DBS surgery for invalidating motor fluctuations. L-Dopa equivalent dose was 1200 mg/day, agonist L-Dopa equivalent dose was 350 mg/day. At this time, he/she presented a marked ‘ON’ hypomanic euphoria, was highly creative and displayed excessive eating behaviour. One year after surgery, the motor effect of STN-DBS was satisfying, allowing for a reduction of L-Dopa therapy to 300 mg/day and dopamine agonist arrest. However, STN-DBS was not satisfying at the psychobehavioural level: irritability with apathy occurred, as well as slight delusions, eating behaviour stayed uncontrolled and risk-taking behaviour appeared. PD dementia was diagnosed 6 years after surgery. Eight years after surgery, behaviour was still difficult to manage: transitory episodes of severe depression, excessive eating behaviour, tantrums, florid hallucinations necessitated all along the follow-up duration to adjust dopaminergic medications, introduce antidepressant and antipsychotic medications, and so on, without success. This case illustrates the ‘oscillating’ profile of patients, particularly hard to manage even with the help of STN-DBS.


This study prospectively described the outcomes of neuropsychiatric symptoms in a cohort of 69 patients with PD treated with bilateral STN-DBS and followed up to 10 years (with a median of 6 years). It expands previously reported data showing a significant reduction of ICDs and neuropsychiatric fluctuations at 1 year after STN-DBS compared with before surgery.21 Specifically, in the same cohort of patients, we have observed at endpoint an overall decrease or complete resolution of hyperdopaminergic symptomatology, namely several behavioural addictions and dopaminergic medication abuse, and a marked decrease in neuropsychiatric fluctuations. However, the retrospective analysis conducted in 48 patients of this cohort has shown that several transient hyperdopaminergic behaviours could occur, namely hypersexuality and excessive eating. Concerning the hypodopaminergic symptomatology, despite medication and stimulation adjustments, apathy was frequently reported in the 69 patients’ follow-up. The retrospective analysis in the 48 patients confirmed and found an even higher occurrence of transient episodes of apathy and depression.

In PD, motor fluctuations are frequently associated with disabling neuropsychiatric fluctuations.29 After STN-DBS, we found a marked improvement in the neuropsychiatric fluctuations that was stable over the years. This important finding is in line with previous work settled with a shorter follow-up.30 Postoperative improvement in neuropsychiatric fluctuations underlies on complex mechanisms, which involve current diffusion to the associative and limbic STN territories (mainly explaining the improvement in off-period dysphoria),31 the decrease in dopaminergic medications and subsequent dopamine receptor desensitisation (mainly explaining the decrease in ON-drug euphoria).32 According to some literature, ICDs are identified in 17.1%–24% of patients with PD treated with dopamine agonists.33 34 The prevalence of ICDs has been linked to the duration of both disease and treatment.6 35 Therefore, patients with PD undergoing STN-DBS have often a neuropsychiatric profile with frequent hyperdopaminergic behaviours.36 We have already shown that bilateral STN-DBS improves hyperdopaminergic symptomatology at 1 year after surgery, and also favours the occurrence of apathy.21 The present study confirms these findings in the long-term follow-up. The beneficial effect is probably due to the reduction of dopamine agonists, known to induce ICDs, and/or to a switch from a pulsatile treatment to a more long-lasting therapy, shared between STN-DBS, apomorphine perfusion and intestinal L-Dopa perfusion.37 38 Interestingly, the proportion of patients presenting excessive eating behaviour remained high in the follow-up. This finding has also been observed in other studies,20 suggesting that the underlying mechanisms of this specific behaviour are more complex and need a different management than just decrease medication.11 20 39 ‘Emotional eating’ in apathetic patients after STN-DBS might favour the onset of pathological eating behaviour,39 and thus explain the observed discrepancy between alleviation of ICDs without decrease in pathological eating behaviour. Surprisingly, among ICDs, only hypersexuality was not significantly alleviated under STN-DBS in our cohort. This finding needs to be confirmed by future prospective studies. An opposite result was observed in a retrospective study of 150 patients treated with STN-DBS for 4 years on average.40 Whereas in prospective studies20–22 STN-DBS allowed improvement in dopamine dysregulation syndrome and behavioural addictions (likely related to the postoperative reductions in dopamine replacement therapy),6 retrospective studies reported no improvement or worsening of pre-existing behavioural addictions, or even new-onset addictions following STN-DBS.23 Our results fit with this apparent paradox of the literature. Indeed, prospective methodology with a single endpoint assessment (evaluating the presence of behavioural symptoms at the time of the evaluation) showed reduction of ICDs whereas retrospective methodology (considering all behavioural symptoms occurred within a long period of time) showed reoccurrence or de novo ICDs in individual patients. These symptoms could have been only transient and thus not detected with the prospective ‘one shot’ analysis. As such, the longer the period of study retrospectively considered, the more pathological behaviours are likely to be found. In our study, we found for a fifth of the patients an oscillating behavioural profile, characterised by alternations of hypodopaminergic and hyperdopaminergic pathological behaviours within the follow-up. The management of these patients is challenging, since the therapeutic window for adapting dopaminergic treatment can be very small: patients can easily switch from apathy to ICDs and vice versa. This demonstrates the complexity of the phenomenon, and the limitations of the preplanned neuropsychological assessment to catch transient symptoms.

Apathy significantly increased in our patients between the preoperative visit and the endpoint visit. Moreover, many patients presented at least one episode of apathy (associated or not with depression) during the follow-up period. Postoperative apathy can be partly explained by the decrease in dopaminergic treatment, and the slowly progressive desensitisation to the psychotropic effects of dopaminergic treatment.32 Besides, treatment-resistant postoperative apathy tends to develop in association with L-Dopa-resistant frontal dysexecutive syndrome, and can be explained by diffuse cortical synucleinopathy rather than STN-DBS direct effects.16 Although postoperative apathy usually responds to dopaminergic treatment,41 apathy remains the most frequent behavioural side effect in the long-term follow-up of STN-DBS. Our group has recently proposed that the occurrence of postoperative apathy in this cohort of patients at 1-year follow-up could counteract the beneficial effects of motor improvement in terms of quality of life.42 In this long-term follow-up, quality of life remained stable despite the progression of PD within the years, suggesting that the long-term management of apathy with dopaminergic drugs was overall beneficial.41 Another explanation may be the occurrence of some long-term adaptation of patients to a more apathetic mode of functioning. Our data moreover show that preoperative hyperdopaminergic behaviours do not lead more frequently to a subsequent apathy.

Seven patients out of 69 (10.1%) developed dementia over a mean follow-up of 6 years. Although a worsening in cognition can be related to surgery and STN-DBS itself,43 it has been mainly attributed to the natural course of the disease.11 Mean disease duration in our cohort was 16 years. Thus, the incidence of dementia in our patients can be explained by the progression of PD.44

The Ardouin Scale and the MINI results about depression showed a trend towards an increase. In the self-questionnaires, this increase reached significance. These results differ from previous randomised controlled trials showing a decrease in self-reported depression scales.10 19 This discrepancy might be partially due to the much longer follow-up of our cohort. Depressive symptoms may appear in case of lack of dopaminergic stimulation,14 and also in reaction to the loss of autonomy associated with disease progression. Indeed, the number of patients under antidepressant therapy increases strikingly in our cohort.

In our study, suicide attempts occurred in 4/69 patients (5.79%) during the 3–10 years’ follow-up, that is, occurred with a prevalence of 0.9% per year. No completed suicide was observed in our cohort. Case reports, and retrospective and observational studies have suggested a striking increase in suicide rate in the first postoperative year,17 but the association between STN-DBS and suicide behaviours remains controversial. A randomised controlled trial suggested that suicide ideation and behaviours were not elevated in the 6-month period post-DBS surgery.18 Both impulsivity and hypodopaminergia may potentially contribute to an increased risk of postoperative suicidal acts.17

At 3 years, when comparing STN and pallidal targets in a large cohort of patients, no pronounced differences in psychiatric outcome were shown.26 Long-term data concerning the behavioural symptoms after pallidal stimulation are not available yet. Therefore, it remains unknown whether different DBS targets have different behavioural outcomes. This information can be very important when addressing the choice of DBS target during the preoperative evaluation of patients with PD. When compared with pharmacological treatment alone in a randomised trial, at 2 years’ follow-up, patients treated with STN-DBS showed a marked alleviation of non-motor fluctuations, an overall better outcome of hyperdopaminergic symptoms, without a significant increased risk of apathy45. It seems mandatory to extend longer such trials in order to disentangle between (1) the effects of STN-DBS with subsequent dopaminergic medication adaptations and (2) the effects of the progression of the disease on these complex psychobehavioural outcomes. In the light of our data, we recommend to inform patients of the postoperative risk of apathy, depression and ICDs: clear explanation of these symptoms is mandatory, and needs to be delivered both to the patient and his relatives. Moreover, it is crucial to explain that these symptoms can occur in the first postoperative year and after, and that they require as soon as possible the involvement of a team specialised in PD since they can be potentially devastating.

This study has some limitations. About 30% of the patients could not be included in long-term assessment; follow-up duration at endpoint visit was heterogeneous among patients; when patients developed dementia, the scales were applied differently then during preoperative assessment (with a close relative); the Ardouin Scale is neither conceived nor validated to be used retrospectively as it was the case here in the retrospective part of the study; we do not mention stimulation parameters such as the total electrical energy delivered or electrode location that can both have important psychobehavioural influence11; comorbidities such as pain, chronic illnesses, and so on, that can influence psychobehavioural symptoms were not recorded. In addition, dementia was recorded according to DSM-IV criteria and not to more recent specific PD criteria.46


Bilateral STN-DBS allows a stable improvement of neuropsychiatric fluctuations in patients with PD. This improvement is durable over several years after surgery. Overall, ICDs are significantly decreased postoperatively. On the other hand, apathy is frequently observed also in the long-term follow-up. Our findings support the safety of STN-DBS also in patients with PD who present with relevant psychobehavioural symptoms and neuropsychiatric fluctuations before surgery. Nevertheless, a close follow-up of such patients is mandatory since at any time in the postoperative period pathological neuropsychiatric events occur: they are transient if detected and treated.


View Abstract


  • MA and EL contributed equally.

  • Contributors Research project: (A) Conception: EL, PK. (B) Organisation: AK, PP. (C) Execution: MA, EL, ST, HK, ESchmitt, AB, JX, VF, AK, ÉSeigneuret, SC, PM, EB, PK. Statistical analysis: (A) Design: EL, PK. (B) Execution: MA. (C) Review and critique: EL, ST, HK, ESchmitt, AB, AC, JX, VF, AK, PP, ÉSeigneuret, SC, PM, EB, EM, PK. Manuscript preparation: (A) Writing of the first draft: MA, EL. (B) Review and critique: MA, ST, HK, ESchmitt, AB, AC, JX, VF, AK, PP, ÉSeigneuret, SC, PM, EB, EM, PK. Final approval of the version published: MA, EL, ST, HK, ESchmitt, AB, AC, JX, VF, AK, PP, ÉSeigneuret, SC, PM, EB, EM, PK. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: MA, EL, ST, HK, ESchmitt, AB, AC, JX, VF, AK, PP, ÉSeigneuret, SC, PM, EB, EM, PK.

  • Funding This research received funding from the Programme Hospitalier de Recherche Clinique Interrégional and Servier Pharmaceutical company NCT01705418.

  • Competing interests ST has received grant from Fondation pour la Recherche Médicale, France Parkinson, and ANR; honoraria, travel grant from Boston, Medtronic, Teva, UCB, AbbVie and Aguettant. AC has received granted funds from Edmond J & Lily Safra Foundation; reimbursement of travel expenses to scientific meetings from Orkyn, Elivie, AbbVie, Medtronic and Boston Scientific. VF has received travel grants from Zambon and AbbVie (World PD congress and MD congress); honoraria from UCB, Medtronic for lecturing; honoraria from AbbVie for participation in scientific committee. JX recieved honoraria and travel grant from Teva, UCB and Aguettant. EL received travel grant from Medtronic. EB reports financial support during the past year outside the submitted work from AbbVie (travel expenses, congress registration, consultant), Medtronic (technical support in our hospital for improving DBS therapy and care of patients with movement disorders), Aguettant (consultant and research project) and public financial support from Agence Régionale de Santé Auvergne et Rhône-Alpes (research project on Parkinson’s disease). HK received reimbursement of registration fees and travel expenses to participate in training session from Medtronic. SC received consultancy fees from Medtronic and Boston Scientific and travel reimbursement from Medtronic and Boston Scientific. ES received travel and congress fees from Medtronic, St Jude and Boston Scientific. PK reports grants from the French Ministry of Health (PHRC Programme de Recherche Hospitalier Clinique Interrégional), and Euthérapie during the conduct of the study; grants and personal fees from Medtronic, Boston Scientific, Movement Disorder Society, UCB; grants from St Jude Medical France, Edmond J & Lily Safra Foundation, INSERM (French National Institute of Health and Research in Medicine), France Parkinson, Swiss National Science Foundation, ROGER DE SPOELBERCH Foundation, Centre National Recherche Scientifique, Orkyn, Homeperf; personal fees from European Society Stereotactic Functional Neurosurgery, outside the submitted work. PM received consultancy fees from Medtronic and congress fees and travel reimbursements from Medtronic and St Jude companies. EM has received honoraria from Medtronic for lecturing and scientific board services. She has also received grant support from Merz.

  • Patient consent Obtained. However, detail has been removed from two of the case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.

  • Ethics approval The Ethics Committee of the Grenoble University Hospital approved the study.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The data belong to the Grenoble University Hospital.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.