Real-world neuropsychiatric safety profile of foslevodopa/foscarbidopa infusion in advanced Parkinson's disease.

BACKGROUND: To describe demographic, clinical, and neuropsychiatric safety outcomes in a cohort of PD patients treated with LDp/CDp infusion in a real-world setting. METHODS: We retrospectively analyzed a database of 77 consecutive PD patients treated with LDp/CDp at two Italian tertiary referral centers, with a minimum 6-month follow-up. Baseline cognitive, epidemiological, and clinical variables were assessed to determine potential neuropsychiatric adverse event (NAE) risk factors. RESULTS: Patients had long disease duration (13.8±6.4 years) and moderate-to-severe motor burden (MDS-UPDRS III: 34.4±15). Pre-existing cognitive impairment (MCI 41.6%, dementia 7.8%), prior hallucinations (26%), and impulse control disorders (27.3%) were frequent. At 6 months, mean L-dopa equivalent daily dose increased by ~300 mg (p<0.001). The dropout rate was 16.9% (13 patients), mainly due to AEs (38.5%), bridging to DBS (23%), or device intolerance (15.4%). New-onset NAEs emerged in 17 patients (22.1%) at a median of 30 days, predominantly hallucinations (41.2%) and psychosis (29.4%). Most events were efficiently managed via careful infusion rate adjustments and low-dose antipsychotics, achieving resolution or improvement in 82% of cases. Three patients (18%) discontinued therapy due to severe NAEs. Multivariable regression analysis confirmed baseline Frontal Assessment Battery (FAB) score as the primary independent predictor of NAE development (OR 0.707, p=0.010). CONCLUSIONS: LDp/CDp infusion is well-tolerated. Although NAEs were more frequent than in pivotal trials, they were largely manageable. Frontal executive dysfunction emerged as a significant risk factor, suggesting specific frontal screening should be routine in candidacy evaluation for cognitively frail patients. Preventive strategies to reduce risk are proposed and discussed.