Atypical Psychosis in Parkinson Disease

Abstract

Background and Objectives

Atypical psychosis, characterized by severe delusions, paranoia, and auditory or somatic hallucinations, is a notable complication of continuous subcutaneous infusion (CSCI) of foslevodopa/foscarbidopa therapy in Parkinson disease (PD). The aim of this study was to identify clinical predictors of CSCI-induced psychosis to understand its potential mechanisms and evaluate predictive measures for early detection and management.

Methods

This retrospective cohort study included patients with PD treated with CSCI (n = 23) and an independent PD database cohort (n = 94) from Osaka University Hospital. In the CSCI cohort, clinical data such as psychosis information and answers from Parkinson's Disease Questionnaire (PDQ39) and the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease–Current Symptoms (QUIP-CS) were collected. Statistical analyses included independent t tests and linear regression to identify predictors of atypical psychosis within a year of CSCI initiation. In the PD database cohort, potential relationships between QUIP-CS scores and other clinical parameters were explored using correlational analyses.

Results

Among the 23 patients, 6 developed atypical psychosis, all occurring within 6 months, with 4 of them discontinuing CSCI. Patients who developed atypical psychosis had significantly higher QUIP-CS scores before CSCI (adjusted p = 0.0032). Linear regression identified QUIP-CS as the sole predictor of atypical psychosis onset (coefficient = 0.199, p < 0.001). Among the PDQ39 subitems, item 27 showed a significant correlation with QUIP-CS scores (r = 0.722, adjusted p = 0.0128). Furthermore, a composite score comprising PDQ39 items 20, 27, 29, 31, and 36 (PDQ39_sub5) showed an even stronger correlation with QUIP-CS scores (r = 0.770, p = 0.0000704). This association was independently confirmed in the PD database cohort (r = 0.415, p = 0.00003). Finally, PDQ39_sub5 effectively stratified survival curves for psychosis onset in the CSCI cohort (p = 0.008).

Discussion

CSCI-induced psychosis is distinct from visual hallucinations observed in typical PD psychosis and likely involves mechanisms in mesolimbic circuits and impulsive-compulsive behaviors associated with dopamine dysregulation. While QUIP-CS is rarely used in clinical practice, widely used PDQ39_sub5 offers a practical way to identify individual psychosis risk. These findings potentially offer tailored strategies to predict and manage atypical psychosis in patients with PD receiving advanced dopaminergic therapies.

Introduction

Parkinson disease (PD) is a progressive neurodegenerative disorder primarily characterized by motor symptoms such as bradykinesia, tremor, and rigidity. These motor impairments are caused by the degeneration of dopaminergic neurons in the substantia nigra, leading to dopamine depletion in the basal ganglia. Dopamine replacement levodopa therapy remains the gold standard for managing motor symptoms.¹ As PD progresses, however, oral levodopa therapy becomes less effective because of its short plasma half-life, leading to motor fluctuations and “off” periods that significantly impair patients' quality of life.² In response to these limitations, advanced therapies have been developed to provide continuous dopaminergic stimulation, reducing fluctuations in dopamine levels.³ Among these, continuous subcutaneous infusion (CSCI) of foslevodopa/foscarbidopa represents a transformative approach. CSCI is a soluble formulation of levodopa and carbidopa prodrugs designed for 24-hour continuous delivery through a minimally invasive subcutaneous infusion system.^4,5 This method offers more stable plasma levodopa and carbidopa concentrations than traditional levodopa therapy, effectively reducing motor fluctuations and improving quality of life.

While CSCI provides significant motor benefits, clinical trials and real-world studies have highlighted psychosis as a notable adverse event associated with its use.^4,6,7 Unlike typical PD psychosis that predominantly involves visual hallucinations, psychosis in CSCI-treated patients is often characterized by auditory and somatic hallucinations, agitation, and delusions.⁸ Similar psychotic symptoms have also been reported in patients receiving high-dose levodopa therapies, such as levodopa-carbidopa intestinal gel therapy, with some studies highlighting the potential involvement of impulse control disorder (ICD) symptoms and dopamine dysregulation syndrome (DDS) in these cases.^9-11 Indeed, CSCI-induced psychosis is speculated to involve sustained stimulation of dopaminergic pathways.⁷ However, further studies are necessary to assess such speculation and understand the potential mechanisms of CSCI-related psychosis.

CSCI-related psychosis underscores the need for careful patient selection and monitoring when initiating CSCI therapy, as well as the development of strategies to mitigate its neuropsychiatric side effects. Therefore, the aim of this study was to elucidate the clinical characteristics and predictors of psychosis in patients with PD receiving CSCI. By analyzing the associations between the onset of atypical psychosis and clinical parameters, we can better understand the mechanisms underlying atypical psychosis associated with CSCI and inform strategies for its management and risk mitigation.

Methods

The protocol conformed to Helsinki Declaration principles and was approved by the Osaka University review board (approval numbers 13471 and 22311). All participants provided written informed consent before any assessment, including the collection of demographic information such as age and sex.

Participants

This retrospective study included 23 patients with PD treated with CSCI at Osaka University Hospital (CSCI cohort) and the patients who participated in a different PD cohort study (PD database cohort) (n = 94) from April 2020 to October 2024. All the patients were confirmed to have clinically established PD according to the Movement Disorder Society's clinical diagnostic criteria, as determined by a movement disorder specialist.¹²

Clinical Assessments

Demographics, clinical characteristics, and questionnaire ratings of the CSCI cohort and PD database cohort are listed in Table 1. The rating scales for PD were assessed by neurologists as previously described.^13,14 To assess the general clinical features of PD, we obtained ratings from the Hoehn and Yahr scale,¹⁵ Movement Disorder Society–Unified Parkinson's Disease Rating Scale,¹⁶ and Parkinson's Disease Questionnaire (PDQ39).¹⁷ Cognitive functions were assessed by the Mini-Mental State Examination (MMSE)¹⁸ and Frontal Assessment Battery. Impulsive and compulsive behaviors were assessed by the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease–Current Symptoms (QUIP-CS).^19,20 Sleep disturbances were assessed by the REM Sleep Behavior Disorder Questionnaire.²¹ Data from the CSCI-treated patients were obtained at the initiation of the treatment, except for the following: QUIP-CS data for patients 9 and 10 were obtained 5 years and 1 year before the CSCI initiation, respectively, and the PDQ39 data for patient 17 was obtained 1 year before the CSCI initiation. Electronic medical records confirmed that these patients' conditions had not changed significantly from the survey.

Table 1.

Demographics, Clinical Characteristics, and Questionnaire Ratings Between Patients Receiving CSCI Treatment (n = 23) and the PD Database Cohort (n = 94)

	CSCI cohort								PD database cohort
	Total		Subgroup
			No psychosis		Typical psychosis		Atypical psychosis
	N	Mean ± SEM	N	Mean ± SEM	N	Mean ± SEM	N	Mean ± SEM	N	Mean ± SEM
Age (y)	23	66.78 ± 1.73	14	66.79 ± 2.21	3	66.00 ± 2.52	6	67.17 ± 4.44	94	70.40 ± 1.02
Age at onset (y)	23	53.22 ± 1.69	14	53.21 ± 2.14	3	53.67 ± 0.67	6	53.00 ± 4.49	94	60.16 ± 1.28
Sex (male: female)	23	11: 12	14	7: 7	3	1: 2	6	3: 3	94	44: 50
Disease duration (y)	23	13.57 ± 1.07	14	13.57 ± 1.39	3	12.33 ± 2.19	6	14.17 ± 2.52	94	10.24 ± 0.69
HY score, disease severity (1: 2: 3: 4: 5)	23	0: 6: 13:: 4: 0	14	0: 3: 8: 3: 0	3	0: 2: 0:1: 0	6	0: 1: 5: 0: 0	94	2: 26: 37: 24: 5
MDS-UPDRS Part III score	23	22.52 ± 1.7	14	22.21 ± 1.75	3	25 ± 9.29	6	22 ± 3.45	94	34.2 ± 1.66
PDQ39 score	23	57.75 ± 4.25	14	54.73 ± 5.72	3	52.67 ± 5.61	6	65.83 ± 9.17	94	53.51 ± 2.21
LED before CSCI (mg)	23	708.17 ± 44.86	14	697.14 ± 60.29	3	616.67 ± 142.4	6	779.67 ± 76.94		N.A.
CSCI initial dose (mL/h)	23	0.27 ± 0.01	14	0.26 ± 0.02	3	0.25 ± 0.06	6	0.29 ± 0.03		N.A.
CSCI dose when psychosis appeared (mL/h)		N.A.		N.A.	3	0.29 ± 0.07	6	0.36 ± 0.04		N.A.
Maximum CSCI dose within 1 y (mL/h)	23	0.33 ± 0.02	14	0.32 ± 0.02	3	0.29 ± 0.07	6	0.38 ± 0.04		N.A.
Agonist use (yes: no)	23	21: 2	14	13: 1	3	2: 1	6	6: 0		N.A.
MMSE score	22	28.23 ± 0.48	13	28.85 ± 0.5	3	27.67 ± 0.88	6	27.17 ± 1.3	94	26.86 ± 0.37
FAB score	20	15.7 ± 0.56	11	16.55 ± 0.47	3	15.33 ± 2.19	6	14.33 ± 1.23	94	14.39 ± 0.35
RBDQ score	19	4.42 ± 0.59	10	4.9 ± 0.67	3	4.67 ± 2.19	6	3.5 ± 1.18	94	4.7 ± 0.33
QUIP-CS score	20	1.5 ± 0.41	11	0.73 ± 0.27	3	0.33 ± 0.33	6	3.5 ± 0.81	94	1.21 ± 0.17

Abbreviations: CSCI = continuous subcutaneous infusion; FAB = Frontal Assessment Battery; HY = Hoehn and Yahr scale; LED = levodopa equivalent dose; MDS-UPDRS = Movement Disorder Society–Unified Parkinson's Disease Rating Scale; MMSE = Mini-Mental State Examination; PDQ39 = Parkinson's Disease Questionnaire; QUIP-CS = Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease–Current Symptoms; RBDQ = Rapid Eye Movement Sleep Behavior Disorder Questionnaire.

Subgroups of the CSCI cohort are as follows: without any psychosis (n = 14), with typical psychosis (n = 3), and with atypical psychosis (n = 6) within 1 y of CSCI treatment.

Survey of Psychosis Onset in CSCI-Treated Patients

Atypical psychosis was defined as auditory/somatic hallucinations and/or delusions with or without visual hallucinations. Typical psychosis was defined as visual hallucinations without auditory/somatic hallucinations. The expert neurologists and the expert psychiatrists confirmed the diagnosis of the psychosis using the electronic medical record. The onset of the atypical psychosis within a year from the initiation of the CSCI was retrospectively surveyed from the electronic medical record.

Statistical Analysis

We used Statistical Package for the Social Sciences 29.0J software (SPSS, IBM Japan, Tokyo, Japan) to perform all statistical analyses. Group comparisons of demographic and clinical variables between patients with and without atypical psychosis were conducted using a two-sample t test with Bonferroni corrections and a χ² test. To identify baseline predictors of CSCI-induced atypical psychosis, a multiple linear regression analysis was performed with MMSE, Frontal Assessment Battery, REM Sleep Behavior Disorder Questionnaire, QUIP-CS, and the self-reported questionnaire for ICD/DDS as explanatory variables. The selection of these explanatory variables was motivated by previously reported risk factors of visual hallucinations and atypical psychosis in PD. Specifically, cognitive impairment and the presence of REM behavior disorder have been associated with the development of visual hallucinations while the presence of ICD and DDS has been associated with atypical psychosis. We used the Kaplan-Meier method to estimate survival curves for the groups divided by the QUIP-CS or by PDQ39_sub5, which was derived from the sum of the 5 PDQ39 subscore items most strongly correlated with QUIP, and the differences were analyzed using the log-rank test.

Bias

Our data are from patients who received CSCI treatment and agreed to participate in this study.

Data Availability

The data sets generated during and/or analyzed during this study are available from the corresponding author on reasonable request.

Results

Clinical and demographic characteristics of the patients in the CSCI cohort and subgroup characteristics—patients without any psychosis (n = 14), those with typical psychosis (n = 3), and those with atypical psychosis (n = 6) within 1 year of CSCI treatment—are listed in Table 1. Clinical data and details of psychosis for each patient are summarized in eTable 1. Among 23 patients, 3 patients developed typical psychosis, which was predominantly characterized by visual hallucination. By contrast, 6 patients developed atypical psychosis, defined by the presence of auditory or somatic hallucinations and/or delusions. The details of the atypical psychosis episode are summarized in eTable 2. For example, somatic hallucinations included “a feeling of the feet being soaked or the body tied with fabric” (case 2) or “the sensation of the drug became electromagnetic fields inside the body” (case 10) while auditory hallucinations included “hearing people in the restaurant talking badly about him” (case 21). An example of a delusion was “the wife joined a cult and has an affair with its leader” (case 7). Specifically, 5 patients experienced auditory hallucinations, 3 had somatic hallucinations, and 3 exhibited delusions. Four of the 6 patients with atypical psychosis also presented with visual hallucinations. Notably, 4 of the patients discontinued CSCI treatment because of atypical psychosis. The onset of atypical psychosis varied among patients, ranging from 9 to 169 days after CSCI initiation (eTable 1).

To understand the clinical parameters that are associated with the onset of psychosis, we compared demographic data and several PD-related clinical scores between the patients with and without atypical psychosis using the Student t tests and χ² analyses where appropriate, followed by the Bonferroni adjustments (Table 2). It is important to note that no patient with typical psychosis terminated the CSCI treatment, indicating the importance of understanding the characteristics of atypical psychosis. Therefore, we pooled the typical psychosis group with the nonpsychosis group to understand specifically the factors associated with the onset of atypical psychosis. We included MMSE, Frontal Assessment Battery, and REM Sleep Behavior Disorder Questionnaire, because cognitive function and REM sleep behavior disorder are known to be associated with conventional visual hallucination^22-24 and may also be related to atypical psychosis. QUIP-CS, a self-reported questionnaire for ICD/DDS, was assessed because ICD/DDS involving the mesolimbic system has been reported to be associated with atypical psychosis.^8,11 There were no differences between patients with and without atypical psychosis in age, sex, disease duration, and severity (p's > 0.05). Their cognitive function and REM sleep behavior disorder symptoms also did not differ (p's > 0.05). There was a tendency for patients with atypical psychosis to have received a higher maximum dose within the first year, although this was not statistically significant (p = 0.165). Among all the variables analyzed, the QUIP-CS score was the only score that showed a significant difference between the 2 groups (p = 0.00021, adjusted p = 0.0032), with a higher score associated with psychosis. We confirmed these results with linear regression analysis, which revealed that QUIP-CS score was the only independent variable associated with the onset of atypical psychosis (Table 3, coefficient = 0.199, p < 0.001), when assessed with other variables known to be associated with psychosis in PD.^25-30 We also stratified the CSCI cohort into 2 groups based on the median split of QUIP-CS scores (QUIP-CS score 0 or 1, n = 12; QUIP-CS score 2 or higher, n = 8). The survival curves for the duration without atypical psychosis for these 2 groups were statistically different according to a log-rank test (Figure 1, p = 0.005), suggesting the predictive value of QUIP-CS scores obtained before CSCI treatment.

Table 2.

Comparisons of the Clinical Data Between Those With and Without Atypical Psychosis Within a Year Since CSCI Treatment

	No atypical psychosis		Atypical psychosis		p Value	Adjusted p
	N	Mean ± SEM	N	Mean ± SEM
Age (y)	17	66.65 ± 1.85	6	67.17 ± 4.44	0.899	1
Age at onset (y)	17	53.29 ± 1.76	6	53.00 ± 4.49	0.471	1
Sex (male: female)	17	8: 9	6	3: 3	0.901	1
Disease duration (y)	17	13.35 ± 1.19	6	14.17 ± 2.52	0.746	1
HY score, disease severity	17	2.94 ± 0.18	6	2.83 ± 0.17	0.742	1
MDS-UPDRS Part III score	17	22.71 ± 2.01	6	22.00 ± 3.45	0.86	1
PDQ39 score	14	54.29 ± 4.57	6	65.83 ± 9.17	0.222	1
LED before CSCI (mg)	17	682.94 ± 54.20	6	779.67 ± 76.94	0.355	1
CSCI initial dose (mL/h)	17	0.25 ± 0.015	6	0.28 ± 0.03	0.325	1
Maximum CSCI dose within 1 y (mL/h)	17	0.32 ± 0.02	6	0.38 ± 0.04	0.165	1
Agonist use (yes: no)	17	15: 2	6	6: 0	0.379	1
MMSE score	16	28.63 ± 0.015	6	27.17 ± 1.30	0.185	1
FAB score	14	16.29 ± 0.56	6	14.33 ± 1.23	0.11	1
RBDQ score	13	4.85 ± 0.67	6	3.50 ± 1.18	0.301	1
QUIP-CS score	14	0.64 ± 0.22	6	3.50 ± 0.81	0.00021	0.0032

Abbreviations: CSCI = continuous subcutaneous infusion; HY = Hoehn and Yahr scale; MDS-UPDRS = Movement Disorder Society–Unified Parkinson's Disease Rating Scale; PDQ39 = Parkinson's Disease Questionnaire; LED = levodopa equivalent dose; MMSE = Mini-Mental State Examination; FAB = Frontal Assessment Battery; RBDQ = Rapid Eye Movement Sleep Behavior Disorder Questionnaire; QUIP-CS = Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease–Current Symptoms.

Chi-squared tests were used for categorical variables (sex and agonist use) while Student t tests were applied to continuous variables.

Table 3.

Linear Regression Analysis to Predict the Onset of Atypical Psychosis Within a Year

	Onset of the atypical psychosis in a y
	Coefficient (95% CI)	p Value
Age	−0.012 (−0.033 to 0.009)	0.225
Disease duration (y)	−0.025 (−0.064 to 0.014)	0.198
HY score	0.053 (−0.250 to 0.35)	0.712
LED (mg)	0.001 (0.00 to 0.001)	0.126
MMSE score	−0.040 (−0.128 to 0.049)	0.351
QUIP-CS score	0.199 (0.098 to 0.300)	< 0.001

Abbreviations: HY = Hoehn and Yahr scale; LED = levodopa equivalent dose; MMSE = Mini-Mental State Examination; QUIP-CS = Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease–Current Symptoms.

Figure 1. Kaplan-Meier Curve of the Onset of Atypical Psychosis Stratified by Median Split of Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease–Current Symptoms (QUIP-CS) Scores Showing Significant Differences in Atypical Psychosis Onset (Log-Rank Test, p < 0.05).

.

QUIP-CS is a measure that is relatively rarely obtained in clinical practice. To find more commonly used clinical scores that can be developed as a new tool for predicting atypical psychosis, we analyzed, in the CSCI cohort, the correlation between QUIP-CS scores and PDQ39 subscores, which is a commonly used assessment measuring a wide range of motor and nonmotor symptoms of patients with PD. Among 39 PDQ39 subscore items, PDQ39 item 27 showed the highest and most significant correlation with QUIP-CS scores (eTable 3, r = 0.722, p = 0.0003, and adjusted p = 0.0128). Following PDQ39 item 27 were items 20, 29, 36, and 31 ranked as top 5 correlations in terms of r value and meeting nominal significance (eTable 3, r > 0.4 and unadjusted p < 0.05). While only the correlation between PDQ39 item 27 and QUIP-CS score survives Bonferroni adjustment, it should be noted that the present unbiased approach of including all 39 subscores likely inflated the type II error rate.³¹ Of interest, when we combined these 5 PDQ39 subscores by adding them together (PDQ39_sub5), the positive correlation with QUIP-CS scores improved compared with any individual score (Table 4, r = 0.77, p = 0.00007, Figure 2A). The content of these 5 questions is further discussed in the Discussion section.

Table 4.

Correlation Between PDQ39_sub5 Score and Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease-Current Symptoms (QUIP-CS) Scores in the CSCI Cohort and the PD Database Cohort

	CSCI cohort (n = 20)		PD database cohort (n = 94)
	r	p Value	r	p Value
PDQ39_sub5* score	0.77	0.00007	0.42	0.00003

Abbreviations: PDQ39 = Parkinson's Disease Questionnaire; PDQ39_sub5= the sum of PDQ39 items 20, 27, 29, 31, and 36.

Figure 2. Evaluation of the Usefulness of PDQ39_sub5.

(A) Correlation of PDQ39_sub5 score and QUIP-CS scores in the CSCI cohort. (B) Kaplan-Meier curve of the atypical psychosis. Stratification by the PDQ39_sub5 score (low: 0–1; high: ≥3) showed significant differences in atypical psychosis onset (log-rank test, p = 0.008). CSCI = continuous subcutaneous infusion; QUIP-CS = Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease–Current Symptoms.

To further confirm the usefulness of PDQ39_sub5 as a potential screening tool to replace QUIP-CS in predicting atypical psychosis, we evaluated the correlation between PDQ39 subscores, PDQ39_sub5 score, and QUIP-CS scores in our PD database cohort (n = 94), which replicated the correlations between PDQ39 items 20, 27, 29, 31, and 36 and QUIP-CS scores (eTable 3). Again, the correlation of PDQ39_sub5 score with QUIP-CS scores was stronger than any individual PDQ39 score (Table 4, r = 0.415, p = 0.00003).

It is important to note that in the CSCI cohort, linear regression confirmed that PDQ39_sub5 independently predicted the onset of the psychosis (Table 5), and the survival curves for atypical psychosis onset divided by the median score of PDQ39_sub5 (low: 0–2; high: ≥3) were statistically different (Figure 2B, p < 0.008). These findings suggest the utility of PDQ39_sub5 as an alternative to QUIP-CS as a predictor of atypical psychosis in CSCI-treated patients, particularly given its accessibility as a routine clinical assessment in PD.

Table 5.

Linear Regression Analysis to Predict the Onset of Atypical Psychosis Within a Year

	Onset of the atypical psychosis in a year
	Coefficient (95% CI)	p Value
	Age	−0.014 (−0.040 to 0.012)	0.259
Disease duration (y)	−0.025 (−0.073 to 0.024)	0.289
HY score	0.060 (−0.312 to 0.433)	0.732
LED (mg)	0.001 (0.00 to 0.002)	0.058
MMSE score	−0.040 (−0.168 to 0.045)	0.235
PDQ39_sub5 score	0.095 (0.021 to 0.169)	0.016

Abbreviations: HY = Hoehn and Yahr; LED = Levodopa Equivalent Dose; MMSE = Mini Mental State Examination; PDQ = Parkinson's Disease Questionnaire; PDQ39_sub5 = the sum of PDQ39 items 20, 27, 29, 31, and 36.

Discussion

While cognitive impairment and advanced age are established risk factors of typical PD psychosis, especially visual hallucinations, our findings suggest a different profile in patients who develop CSCI-induced atypical psychosis, characterized by auditory and somatic hallucinations, agitation, and delusions. In our cohort, MMSE and Frontal Assessment Battery (FAB) scores were not significantly different between groups, whereas QUIP-CS score, which is a screening tool designed to assess impulsive and compulsive behaviors and DDS in patients with PD,²⁰ showed a strong association. These findings align with previous studies suggesting the existence of a unique type of psychosis in PD characterized by symptoms of ICD and DDS that are distinct from typical PD psychosis predominantly involving visual hallucinations.¹¹ This unique type of PD psychosis is less associated with cognitive dysfunction or older age, and the symptoms often involve mania-like or schizophrenia-like delusions and auditory/somatic hallucinations.¹¹ Notably, the significantly higher QUIP-CS scores measured before CSCI therapy in patients with atypical psychosis suggest that compulsive medication use and ICD-like behaviors were present before the onset of psychotic symptoms.

The psychosis observed in our CSCI-treated patients also shares similarities with the concept of hedonistic homeostatic dysregulation (HHD), a syndrome described in patients with PD undergoing dopamine replacement therapy.³² HHD is characterized by compulsive dopaminergic medication overuse, mood dysregulation, ICD-like behaviors, and manic psychosis.³² A defining feature of HHD is compulsive drug-seeking behavior often in the absence of a clear motor need, leading to excessive dopamine in the CNS. In HHD, patients experience withdrawal-like dysphoria or irritability when medication is reduced, reinforcing compulsive drug seeking.³² Similarly, continuous high-dose dopaminergic stimulation through CSCI may exacerbate dysregulated reward and salience processing in predisposed patients with ICD symptoms to result in paranoid delusions, auditory/somatic hallucinations, and agitation.

Given that HHD, DDS, and ICD all involve dysfunction in the mesolimbic dopamine system,^32-34 it is likely that CSCI-induced psychosis stems from similar neurobiological mechanisms. This supports the hypothesis that atypical psychosis in CSCI-treated patients with PD is linked to excessive dopamine stimulation in the nucleus accumbens (NAc), a structure strongly implicated in HHD and DDS.^35,36 Indeed, recent neuroimaging studies, including our own work, have demonstrated that functional connectivity between the NAc and cortical regions is altered in PD patients with ICDs and DDS.³⁷ Specifically, reduced connectivity between the NAc and ventromedial prefrontal cortex has been associated with impulsive and compulsive behaviors, potentially contributing to the heightened risk of DDS observed with high-dose continuous dopamine delivery. Given that the ventral striatum, particularly the NAc, plays a central role in reward processing and salience attribution, it is plausible that aberrant salience processing within mesolimbic circuits may underlie the delusions, auditory/somatic hallucinations, and paranoia observed in CSCI-treated patients.^38,39 We hypothesize that the QUIP-CS score predicts atypical psychosis in CSCI-treated patients with PD because of its ability to capture NAc-related dysfunction, which warrants further investigation in future studies.

Moreover, the identification of PDQ39_sub5 as a significantly and positively correlated measure with QUIP-CS is particularly interesting. It comprises the following 5 items: Q20: Felt angry or bitter? Q27: Had problems with your close personal relationships? Q29: Lacked support in the ways you need from your family or close friends? Q31: Had problems with your concentration, e.g., when reading or watching TV? Q36: Felt ignored by people? These questions evaluate the deficits in emotional and attentional control and the disruption of the relationship with others,¹⁷ which mimic symptoms of manic psychosis, providing converging evidence of mesolimbic dopaminergic system dysfunction in CSCI-induced psychosis. The strong correlation between PDQ39_sub5 and QUIP-CS suggests that specific aspects of emotional regulation and interpersonal relationships may be particularly relevant in the development of CSCI-induced psychosis. Given that PDQ39 is widely used in routine clinical practice, these subscores provide a practical and accessible tool for identifying patients at higher risk of developing atypical psychosis. That is, PDQ39_sub5 could serve as an alternative screening measure for atypical psychosis risk before CSCI initiation, which may facilitate earlier intervention and improved clinical management.

The pharmacokinetic challenges of continuous dopaminergic therapies also warrant discussion. Before the introduction of CSCI, studies on levodopa-carbidopa intestinal gel have demonstrated both benefits of motor symptom management and risks of neuropsychiatric complications. While continuous delivery mitigates motor fluctuations, the sustained stimulation of dopaminergic pathways may exacerbate predispositions to manic psychosis. Case reports of DDS in patients receiving levodopa-carbidopa intestinal gel resembled what we experienced in CSCI, suggesting a potential shared mechanism of increased neuropsychiatric side effects in continuous dopaminergic therapies.^9,10 In addition to the effect of continuous delivery, increased dosage may also contribute to the development of CSCI-related psychosis. Symptoms such as structured hallucinations and delusions have been associated with higher doses of dopaminergic drugs.⁸ In our study, patients who developed atypical psychosis tended to have received higher doses within the first year, although the difference was not statistically significant. Based on our clinical observations, this may reflect a scenario in which patients with DDS experience frequent “no-on” periods after switching to CSCI, prompting dose escalation, which in turn increases the risk of psychosis. Meanwhile, the reduced risk of motor complications such as dyskinesia may make clinicians more willing to increase the dose when patients report persistent off periods. These findings further highlight the intricate balance between achieving motor control and avoiding neuropsychiatric side effects, especially in patients with DDS.

There are some limitations in this study. The utility of QUIP-CS and PDQ39_sub5 in predicting the onset of atypical psychosis needs to be prospectively confirmed in future studies. In addition, it is yet unclear what factors drive some patients to terminate CSCI while the others do not. Nevertheless, this study offers a practical way to promote precision medicine in PD. A balanced approach would be to ensure that patients identified to have a high risk of atypical psychosis with CSCI therapy are not excluded from such transformative therapy but instead receive special attention, such as close monitoring for early signs of psychosis and the implementation of a secondary treatment plan to manage it.

This study highlights the atypical nature of CSCI-induced psychosis in PD and its association with ICD/DDS. The correlation between QUIP-CS and PDQ39_sub5 underscores the utility of these measures in clinical practice. Our findings suggest that psychoses in PD encompass distinct subtypes with potentially different neurobiological mechanisms, warranting tailored therapeutic approaches.

Author Contributions

L. Ge: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. Y. Kimura: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. K. Kakuda: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. K. Ogawa: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. Y. Kajiyama: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. K. Asai: major role in the acquisition of data. S. Taniguchi: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. G. Beck: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. Y. Nishio: drafting/revision of the manuscript for content, including medical writing for content. J.H. Kim: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. K. Ikenaka: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. H. Mochizuki: drafting/revision of the manuscript for content, including medical writing for content.

Study Funding

This research was supported by Japan Agency for Medical Research and Development under Grant Number JP25gm1910008h, JST Fusion Oriented Research for disruptive Science and Technology Program under Grant Number JPMJFR231L, and Japan Society for the Promotion of Science KAKENHI under Grant Number JP24H00045.

Disclosure

K. Ikenaka has received speaker's honoraria from Eisai, FP, and AbbVie. L. Ge, Y. Kimura, K. Kakuda, K. Ogawa, Y. Kajiyama, K. Asai, S. Taniguchi, G. Beck, Y. Nishio, J.H. Kim, and H. Mochizuki report no disclosure. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.