Parkinson’s disease (PD) is a neurodegenerative disorder characterized by motor symptoms, including bradykinesia, rigidity, tremor, and postural instability [1]. Patients with advanced-stage PD often suffer from motor fluctuations due to a narrowed therapeutic window, and some of these patients are suitable candidates for device-aided therapies (DATs) [2]. The newly developed soluble prodrug formulation of levodopa (LD)/carbidopa (CD), foslevodopa/foscarbidopa, is administered as a 24-h/day continuous subcutaneous infusion (CSCI) to manage motor fluctuations in advanced-stage PD. In a phase III trial, CSCI significantly reduced the duration of motor fluctuations, whereas 26.2% of PD patients discontinued CSCI because of adverse events, with hallucination being the most common (4.1%) [3]. Although the exact underlying mechanism remains unclear, hallucinations can be induced by the continued use of dopamine agonist therapy and unexpected increases in plasma LD concentrations [4,5]. It is necessary to define appropriate PD targets, establish practical ways to initiate CSCI, and clarify the optimal role of CSCI compared with other DATs (e.g., deep brain stimulation [DBS] and LD/CD intestinal gel [LCIG]) [6]. Here, we report the case of a 56-year-old woman whose treatment was switched from LCIG to CSCI, but the CSCI was eventually stopped because of psychosis caused by unexpected increases in plasma LD concentrations.

The patient was diagnosed with PD at 39 years of age, and the patient’s first symptom was a tremor in her left upper limb. Motor symptoms responded well to anti-parkinsonian medications for 7 years, but the PD slowly progressed. Wearing off and dyskinesia emerged at the ages of 46 and 47 years, respectively. As motor fluctuations became more pronounced, treatment with 100 mg of LD (administered 5–8 times/day) or 13.5 mg of the rotigotine transdermal patch was used. At the age of 50 years, a treatment switch to DATs, including DBS and LCIG therapy, was proposed. The patient declined to undergo intracranial surgery and chose to start LCIG treatment. The LCIG effectively reduced the occurrence of the patient’s motor fluctuations. Although she transiently experienced visual hallucinations with LCIG, these were ameliorated by decreasing the LCIG infusion rate and dopamine agonist doses. During the use of LCIG, the patient frequently suffered from dermatological issues at the gastrostomy site and tube-associated issues, including tube malpositioning, leading to a switch from LCIG to CSCI when the patient was 56 years old. At that time, cognitive function tests revealed normal scores and findings: Mini-Mental State Examination (MMSE) score, 27; Montreal Cognitive Assessment Japanese version (MoCA-J) score, 26; and noise pareidolia test, negative findings. On an LD challenge test (LCT), the rate of improvement in the Movement Disorders Society-Unified Parkinson Disease Rating Scale (MDS-UPDRS) Part III score was 38.4% (MDS-UPDRS Part III scores: 32 in the ON-state, 52 in the OFFstate). Serial blood samples were collected, and plasma LD concentrations during the LCT (Figure 1A) and following the initiation of LCIG (without a morning dose) (Figure 1B) were determined by high-performance liquid chromatography. Analysis of the pharmacokinetic data revealed that the LD concentration peaked within 15–30 min after 150 mg of LD oral intake (her usual morning oral LD dose, 100 mg) [7] and decreased to less than 1,000 ng/mL at 2 hours post-LD intake. During the LCIG treatment, the LD concentration increased to more than 2,000 ng/mL at 1 hour after the introduction of LCIG and stabilized at 2,000–3,000 ng/mL. The LCIG dose was equivalent to a 1,120 mg levodopa equivalent daily dose (LEDD) (LCIG infusion rate of 3.0 mL/h, daily usage time of 16 hours, morning dose of 8 mL, and drug concentration of 20 mg/mL, without considering the conversion coefficient of intrajejunal LD/ CD infusion; 1.11) [8], and CSCI was started at a 0.37 mL/h infusion rate (62.9 mg/h as LD/CD), with a lower alternative infusion rate of 0.15 mL/h (25.5 mg/h as LD/CD) (day 1). The only concomitant medication used was a 6.25 mg rotigotine transdermal patch. At 4 hours after the start of CSCI, the patient’s dyskinesia became more severe; the infusion rate was reduced to 0.34 mL/h, alleviating dyskinesia. Plasma LD concentrations were determined after CSCI initiation (without a loading dose) (Figure 1C). Analysis of the pharmacokinetic data revealed that the LD concentration continued to increase to 2,900 ng/mL at 6.5 hours after the start of CSCI, although the infusion rate had been decreased 2.5 hours before. Additional CSCI doses or oral LD intake were used when motor symptoms worsened. However, bradyphrenia and psychosis, including visual hallucinations, persecutory delusions, and suicidal ideation, were observed irrespective of motor symptom severity (day 5). On day 7, the CSCI stopped at night, and the psychosis improved the next morning. CSCI was thus restarted with a lower infusion rate of 0.25 mL/h. After the CSCI restart, the patient remained mentally unstable, and her motor symptoms did not sufficiently resolve, resulting in the discontinuation of another CSCI (day 8). The plasma LD concentration during psychosis, including delusions, was 4,100 ng/mL (with a CSCI infusion rate of 0.34 mL/h, 2.5 hours after 100 mg of LD intake), suggesting that the LD concentration might continue to increase for a few days after the introduction of CSCI and cause LD-induced psychosis. After CSCI discontinuation, LCIG was restarted, and the patient’s motor and non-motor symptoms were promptly alleviated.

Foslevodopa/foscarbidopa can achieve continuous dopaminergic stimulation and offers significant benefits for motor function and motor complications in PD patients [9]. For this patient, treatment was switched from LCIG to CSCI per the provided conversion algorithm [4]. As the conversion coefficient of intrajejunal LD/CD infusion was not considered, the true LEDD of LCIG (including the 8 mL morning dose) was not 1,120 mg but was 1,243.2 mg, meaning that the LEDD of CSCI was lower than that of LCIG when the treatment was switched from LCIG to CSCI. However, even with a lower LEDD, this switch resulted in psychosis due to unexpected high LD plasma concentrations. Differences between drug delivery systems, particularly in terms of intestinal absorption versus subcutaneous infusion, can lead to unexpected increases in LD concentrations during CSCI therapy [5]. It is crucial to determine the optimal infusion rate gradually, on the basis of the patient’s clinical condition, rather than relying solely on conversion algorithms. Moreover, as plasma LD concentrations were measured only for 6 hours after the initiation of CSCI, it is important to recognize that the stabilization of plasma LD concentrations may take longer. In fact, a previous clinical trial revealed that the serum pharmacokinetics of CSCI increased more slowly to reach a plateau than did those of LCIG [10], so it would be safer to start CSCI at a lower infusion rate than that calculated from the LEDD of LCIG. This patient may be predisposed to the cognitive and psychiatric side effects of CSCI, given her history of visual hallucinations during LCIG therapy, despite normal results on cognitive tests, including the MMSE, MoCA-J, and noise pareidolia test.

In conclusion, a CSCI protocol of setting a lower infusion rate in the initiation phase, monitoring the patient’s condition and adjusting the infusion rate carefully over the long term may minimize the risk of unexpected increases in plasma LD concentrations during this therapy. Further studies are necessary to establish safe and effective protocols for initiating CSCI.

Citations to this article as recorded by