Introduction

Parkinson´s disease (PD) is a complex neurodegenerative disorder that often presents with a variety of non-motor symptoms, especially neuropsychiatric symptoms1. People with PD might report anxiety, sadness, fatigue, lack of energy and motivation when they are in the OFF-medication (hypodopaminergic) condition, and euphoria, well-being, impulse control disorders (ICD), hypomania, and psychosis under the ON-medication (hyperdopaminergic) condition2,3. However, neurotransmitters other than dopamine (e.g., serotonin, norepinephrine, acetylcholine, and adenosine) also contribute to various neuropsychiatric states1. Additionally, non-motor symptoms frequently fluctuate3,4 both in parallel with, and independently from, motor fluctuations5.

Identification and treatment of both neuropsychiatric symptoms and fluctuations should be part of the routine patient management6. Several tools are available to detect both non-motor symptoms7,8 and non-motor fluctuations9,10, as well as a behavioral scale for the assessment of mood fluctuations11. However, all these instruments are retrospective, and none enables specific and acute assessment of neuropsychiatric fluctuations. To fill this gap, we developed a self-administered scale, the Neuropsychiatric Fluctuations Scale (NFS), specifically designed to be completed during different acute medication conditions (thus not retrospectively)2. The NFS has been shown to be effective in identifying and quantifying acute neuropsychiatric symptoms in both the OFF- and ON-medication conditions12,13. Moreover, a single administration in patients without neuropsychiatric fluctuations could capture chronic hypo- or hyperdopaminergic state2.

Similar to motor fluctuations, neuropsychiatric fluctuations might respond to oral dopaminergic treatment optimization3,5. Advanced treatments such as subcutaneous apomorphine (APO) and levodopa/carbidopa intestinal gel (LCIG) pumps may be effective for non-motor symptoms, but they have not been specifically investigated for the treatment of neuropsychiatric fluctuations14. Bilateral subthalamic nucleus (STN) deep brain stimulation (DBS) can also effectively reduce the overall burden of non-motor symptoms15. Post-operative improvement of health-related quality of life was notably correlated to the improvement of neuropsychiatric symptoms such as mood problems, apathy, attention, and memory changes16. Furthermore, STN-DBS has been reported to reduce neuropsychiatric fluctuations17, likely even better than the best medical therapy alone18. However, all these results are based on retrospective data, thus with possible related biases.

The main objective of the present study was to acutely assess the impact of STN-DBS on neuropsychiatric fluctuations in people with PD using the NFS before and 1 year after surgery.

Results

A total of 48 PD patients were included in the study. One patient could not complete the NFS under the ON-medication condition prior to surgery (spell of discomfort), and data were missing in the OFF-medication condition one year after surgery in two patients for technical reasons. Table 1 shows the main demographic and clinical characteristics before and after STN-DBS for the 45 patients with complete data set.

Table 1 Patient demographic and clinical characteristics at baseline and one-year after STN-DBS
Full size table

Results are given as mean ± SD. At 1-year after surgery, there was a significant improvement in the MDS-UPDRS part III score in the OFF-medication/ON-stimulation condition, in the MDS-UPDRS parts I, II, and IV scores in the ON-medication/ON-stimulation condition, and in the ASBPD compared to before surgery (Table 1). On the contrary, there was a worsening of the Frontal Score after surgery (Table 1). The reduction of LEDD (about 52%) was also significant.

Primary outcome

Neuropsychiatric fluctuations as measured by the TFS were significantly reduced after surgery (30.7 ± 13.7 before surgery vs. 18.7 ± 12.8 one year after surgery; t = 6.22; p < 0.001). As can be seen in Fig. 1, the amplitude of the fluctuations, i.e., the difference in the NFS global scores between the ON- and OFF-medication conditions, was greater before surgery than after surgery.

Fig. 1: Left axis, diamonds: The effects of surgery on the Total Fluctuation Score.
figure 1

Right axis, circles: The effect of medication and surgery on the Neuropsychiatric Fluctuation Scale global score. Data are presented as mean ± standard error of mean. *p < 0.001. NFS, Neuropsychiatric Fluctuations Scale.

Full size image

Secondary outcomes

The effects of both Medication (p < 0.001) and Time point (p = 0.003) on the NFS-global score were significant, as was the interaction between the two factors (p < 0.001). The NFS-global score decreased under medication, and also decreased after compared to before surgery. Examination of the interaction showed that the reduction in the NFS-global score one year after surgery compared to before surgery was significant only in the OFF-medication condition (OFF-medication: −15.24 ± 10.77 vs. −5.42 ± 12.02; p < 0.001; ON-medication: 15.49 ± 9.43 vs. 13.29 ± 8.93; p = 0.158) (Fig. 1).

Comparison of the NFS sub-scores revealed a significant decrease in the NFS-minus sub-score one year after surgery compared to baseline in the OFF-medication state only (p < 0.001). Regarding the NFS-plus sub-score, it was significantly increased after surgery compared to before in the OFF-medication condition (p = 0.001), and significantly decreased in the ON-medication condition (p = 0.013) (Fig. 2).

Fig. 2: Changes in the NFS-minus (left) and NFS-plus (right) sub-scores in the two medication conditions before and after surgery.
figure 2

Data are presented as mean ± standard error of mean. NFS Neuropsychiatric Fluctuations Scale, Med Medication, Post-op one year after surgery. *Statistically significant difference.

Full size image

A significant correlation was found between the NFS-global score and the MDS-UPDRS part III score in the OFF-medication condition before surgery (rho = −0.572; p < 0.001), but not at 1-year follow-up (r = −0.105; p = 0.490).

The correlation between the NFS-global score and the MDS-UPDRS IV score in the OFF-medication condition was significant before (rho = −0.407; p = 0.006) but not after surgery (rho = −0.246; p = 0.103).

There was no correlation between the change in the TFS and the change in LEDD (r = 0.233; p = 0.123) after surgery. Moreover, at 1-year follow-up, there were no correlations between the NFS-plus sub-score in the ON-medication condition and LEDD (r = 0.075, p < 0.624), the change in LEDD (r = 0.070, p < 0.647), and the percent reduction of LEDD (r = 0.093, p < 0.543).

Discussion

Using a specifically designed scale, the NFS, in a well-defined cohort of 45 PD patients, we found a beneficial effect of bilateral STN-DBS on reducing neuropsychiatric fluctuations at 1-year after surgery. Indeed, there was a significant reduction of the NFS-global score in the OFF-medication condition, and a marginal one in the ON-medication condition, leading to a 42% reduction of neuropsychiatric fluctuations one year after surgery.

As already shown in previous large randomized trials19,20, the scores of all the MDS-UPDRS parts improved after surgery, except for the motor score in the ON-medication condition which remained stable despite a significant reduction of LEDD.

Regarding correlation analyses, the significant negative correlation between the NFS-global score and the MDS-UPDRS parts III and IV existing in the OFF-medication condition at baseline, was no longer present one year after surgery. Finally, the change in TFS was not linked to the change in LEDD.

This is one of the few studies acutely assessing the impact of STN-DBS on neuropsychiatric fluctuations using a specific neuropsychiatric scale. Over the past few years, research has been mostly focused on non-motor symptoms or non-motor fluctuations in general15,21,22. A recent combined retrospective-prospective controlled study15 did not find significant improvement in neuropsychiatric symptoms when comparing PD patients treated with STN-DBS to patients treated with best medical treatment at 3-year follow-up. However, to assess non-motor symptoms the authors used the retrospective Non-Motor Symptoms Scale (NMSS)23 which includes only two neuropsychiatric symptoms, namely mood/apathy and perceptual problems/hallucinations. Moreover, the study did not specifically address non-motor fluctuations.

Few data are available about the specific effect of STN-DBS on neuropsychiatric fluctuations17,18,24. In an earlier study that prospectively investigated behavior in 63 PD patients before and 1-year after bilateral STN-DBS using the ASBPD, we reported improvement in mood fluctuations in both the OFF-medication and ON-medication states24. More recently, using again the ASBPD in 69 PD patients with at least 6-year follow-up after STN-DBS, we found a significant improvement on neuropsychiatric fluctuations (namely for euphoria and dysphoria) compared to baseline17. However, the ASBPD is a retrospective hetero-assessment using a semi-structured interview, and it has been shown that hetero-assessments are less reliable than self-questionnaires7. In addition, the ASBPD has only two items for neuropsychiatric fluctuations: ON-medication euphoria and OFF-medication dysphoria. Therefore, the positive outcomes of these studies were awaiting additional support regarding neuropsychiatric fluctuations. In a cohort of 20 PD patients with STN-DBS, another group reported some benefit in non-motor fluctuations, including psychiatric fluctuations, only in the OFF-medication condition at 2-year follow-up22. The authors used a custom-made retrospective questionnaire (based on the one previously used by Witjas et al. 25) that consisted of 25 items across four categories, including psychiatric symptoms, and they also created a scale to record the severity of non-motor fluctuations.

Compared to the above studies, our findings are specific to neuropsychiatric fluctuations, and they are not biased by the retrospective nature of the scale used to assess the outcomes. This point is critical especially considering that significant changes in cognitive tests are frequently found after surgery26. Indeed, worsening in attention and memory can impact the reliability of recalling symptoms.

The improvement of the neuropsychiatric fluctuations after surgery is likely at least partly linked not only to the improvement of motor function but also to the decrease of motor complications. Indeed, the correlation between the NFS-global score and the MDS-UPDRS part III and part IV scores in the OFF-medication condition observed at baseline were no longer significant one year after surgery. Moreover, changes in the LEDD after surgery were not linked to the changes of neuropsychiatric fluctuations. These findings are intriguing and suggest a key role of STN stimulation in modulating neuropsychological symptoms and fluctuations. Indeed, the direct impact of STN stimulation on specific non-motor symptoms in acute and long-term setting is known27. The pathophysiological bases of non-motor symptoms are related to the disruption of different neurotransmitter systems (notably dopamine, but also other systems) due to PD neurodegeneration28, and it is hypothesized that symptom-specific neural network changes play an important role29. DBS can alleviate or aggravate neuropsychiatric symptoms via modulating the non-motor parts of the STN, current spreading to neighboring areas, or modulating the basal ganglia-thalamo-cortical loops29.

Interestingly, contrary to what occurred in the OFF-medication condition, we did not find a significant improvement in the NFS-global score after surgery in the ON-medication condition. Thus, our study suggests that STN-DBS reduces neuropsychiatric fluctuations between the two medication conditions mainly through the reduction of hypodopaminergic symptoms in the OFF-medication condition, whereas there is a more limited reduction of the hyperdopaminergic symptoms in the ON-medication condition. Indeed, the observed reduction in the NFS-plus sub-score at follow-up could be due to the reduction of dopaminergic treatment rather than to a direct effect of STN-DBS. We did not find any correlation between LEDD reduction and NFS-plus sub-score in ON-medication after surgery. In other words, the remarkable LEDD reduction allowed by surgery can be translated into reduced hyperdopaminergic sub-scores, i.e., a normalization of the neuropsychiatric state of patients, rather than a deficit in compensation of dopa-responsive symptoms. However, we cannot exclude a possible impact of the stimulating contact of the DBS lead, since the effects of STN-DBS on different non-motor symptoms has been reported to be linked to the stimulation of STN sub-regions (e.g., with better effect on apathy with the electrode placed more ventrally to the STN and worse effect with the usual dorsolateral placement)29.

Our study also showed a temporal connection between motor function and neuropsychiatric state in the OFF-medication condition at baseline, as already shown in earlier research12,30. These findings can be explained by the hypodopaminergic background of these two entities31. However, this bond was lost after the surgery as discussed above.

Our study has some limitations. First, the sample size is limited. During the COVID-19 pandemic, the number of DBS surgeries was greatly reduced in our center, with the cancellation of all non-vital surgeries, preventing inclusion of a greater number of patients in the present study. Second, the NFS is rather a new scale. However, results using the NFS have already been published2,12,32, including a pre-validation study13. Moreover, the NFS has been designed to quantify the overall neuropsychiatric symptoms and to measure their fluctuations between different medication states. As such, it cannot be used to distinguish different symptoms individually.

To conclude, this study has allowed to unveil the specific impact of STN-DBS on neuropsychiatric fluctuations. Previous studies were scarce and unspecific, especially because the lack of tools enabling to detect these fluctuations accurately and acutely. Here, we have used the NFS, a specific non-retrospective scale, to directly measure neuropsychiatric fluctuations, as well as neuropsychiatric symptoms in both ON- and OFF-medication conditions. We have shown that bilateral STN-DBS effectively reduces neuropsychiatric fluctuations.

The NFS promises to be a useful tool not only in the research settings, but also in the practical clinical work.

Methods

Population

Consecutive patients with PD (according to the criteria proposed by the International Movement Disorders Society, MDS)33 who received STN-DBS (using the previously described criteria)34 at the Movement Disorders Center of the Grenoble Alpes University Hospital (Grenoble, France) between 2016 and 2021 were included.

Inclusion criteria were bilateral STN-DBS surgery, and the preoperative and 1-year postoperative completion of the NFS. Exclusion criteria were patients with diagnosis other than PD, and patients with unilateral STN-DBS or different DBS targets.

The local ethical committee approved the study (NCT04608123). All patients gave their informed consent.

Assessments

All included patients underwent preoperative and 1-year postoperative standardized neurological and neuropsychiatric evaluations.

Neurological assessment included the MDS Unified Parkinson´s Disease Rating Scale (MDS-UPDRS) for rating cognitive and psychiatric symptoms (part I), activities of daily living (part II), motor symptoms of PD (part III), and motor complications (part IV)35.

Neuropsychiatric fluctuations were assessed using the NFS2. Briefly, the NFS consists of 20 items: ten NFS-plus and ten NFS-minus items, corresponding to neuropsychiatric symptoms usually linked to hyperdopaminergic and hypodopaminergic states, respectively.

The NFS provides two sub-scores, a NFS-plus sub-score corresponding to positive (or hyperdopaminergic) symptoms, and a NFS-minus sub-score corresponding to negative (or hypodopaminergic) symptoms, each with a maximum of 30 points. The NFS global score is calculated by subtracting the NFS-plus sub-score from the NFS-minus sub-score, thus ranging from −30 to +30. The more negative the score, the more the patient experiences hypodopaminergic symptoms. Conversely, the more positive it is, the more the patient experiences hyperdopaminergic symptoms. Importantly, the NFS allows to directly assess the amplitude of the fluctuations between the OFF- and ON-medication conditions, using a total fluctuation score (TFS). The TFS is calculated by subtracting the NFS-global score obtained under the ON-medication condition from the NFS-global score measured during the OFF-medication condition (NFS-global ON – NFS-global OFF), ranging from 0 to 60. The closer to 60, the greater the amplitude of the fluctuations between the two medication conditions.

Neuropsychiatric evaluation also included the following tests: the Beck Depression Inventory (BDI-II) for assessing depression, the Starkstein Apathy Scale (SAS) for assessing apathy, the neuropsychiatric fluctuations part of the Ardouin Scale of Behaviour in PD (ASBPD) for evaluating ON-state euphoria and OFF-state dysphoria36, the Mattis Dementia Rating Scale (MDRS-2) and the Frontal Score for comprehensive evaluation of cognition.

Before surgery, both the MDS-UPDRS part III and the NFS were completed in the “defined-OFF” (at least twelve hours after receiving the last dopaminergic treatment dose), and in the “defined-ON” conditions (about 45–60 min after an acute levodopa challenge) conditions37. At 1-year follow-up, the two scales were administered in both the OFF-medication and ON-medication conditions, and with stimulation ON (with the chronic parameters of stimulation).

The total daily levodopa equivalent dose (LEDD) was calculated for all the patients before and after surgery38.

Statistical analysis

Descriptive statistics were used to analyze data regarding demography and disease duration, dopaminergic medication, and assessment results. Continuous variables were expressed as mean and standard deviation ( ± SD). Differences in the MDS-UPDRS parts I-IV, the BDI-II, the SAS, the ASBPD, the MDRS-2, the Frontal Score, and the LEDD between before and after surgery were analyzed using parametric tests (paired t-test) for normally distributed data, and non-parametric tests (Wilcoxon) otherwise.

The primary outcome was the difference in the TFS between before and one year after surgery, assessed using a paired-sample T-test. A p-value < 0.05 was considered significant.

Secondary outcomes were: 1) the difference in the: a) NFS-global score and b) NFS-minus and NFS-plus sub-scores between before and 1 year after surgery, in both the OFF-and ON-medication conditions. Scores were compared using a 2 (Medication, OFF vs. ON) x 2 (Time point, pre-surgery vs. one year after surgery) repeated measures ANOVA. Interactions between the two factors were examined using Holm post-hoc tests; 2) the correlations between the NFS-global score and: a) the MDS-UPDRS part III score in the OFF-medication condition before and 1 year after surgery; b) the MDS-UPDRS part IV score (motor complications) in the OFF-medication condition before and 1 year after surgery; 3) the correlation between the change in the TFS score and the change in LEDD at 1-year follow-up; 4) the correlation between the NFS-plus sub-score in the ON-medication condition, the absolute LEDD, and the change in LEDD at 1-year follow-up.

The assumption of normality was assessed with the Shapiro-Wilk test. Pearson’s correlations were computed for normally distributed data and Spearman´s rank-order correlations were computed for not-normally distributed ones. For all statistical analyses, p-values < 0.05 were considered statistically significant.

Statistical analyses were performed using JASP (version 0.15; JASP team 2021, University of Amsterdam).