Prospective longitudinal assessment of hallucinations in Parkinson's disease

To monitor the evolution of hallucinations over 4 years in a stratified sample of patients with PD.

Using a modified version of the Unified PD Rating Scale (UPDRS) Thought Disorder question, the authors stratified patients into five baseline behavioral groups. They recruited up to 20 patients for each group to participate in sequential interviews (Rush Hallucination Inventory) at baseline and 6, 18, and 48 months. UPDRS motor examinations and Mini Mental State Examinations (MMSE) were obtained at baseline and 48 months. Data were analyzed with Wilcoxon rank sum tests, Mantel-Haenszel tests, and Spearman correlations. To determine features that influenced the new development of hallucinations, a cumulative logit regression model of hallucination severity over time was fit using generalized estimating equations.

Based on the design stratification, 60 patients had no hallucinations at baseline (20 with no behavioral problems, 20 with sleep fragmentation, 20 with altered dream phenomena). Twenty-nine patients had hallucinations (20 with retained insight and 9 with loss of insight). At 48 months, the authors could account for all but two subjects (98% retrieval). In 4 years, the presence of hallucinations increased (33% at baseline, 44% at 18 months, and 63% at 48 months, p < 0.0001). The presence of frequent hallucinations (at least three times weekly) also increased (p = 0.0002). Having hallucinations at baseline or at any given assessment was a strong predictor at all follow-up evaluations of continued hallucinations (p < 0.0001). Hallucinations were not associated with increased mortality (chi(2) = 0.59, df (1), p = 0.47). Among the 60 subjects without hallucinations at baseline, time was the only significant factor influencing the development of hallucination over 48 months. Baseline age, PD duration, sex, medications, and UPDRS or MMSE scores did not influence the incidence of hallucinations.

This prospective, longitudinal study documents the persistent and progressive nature of hallucinations in PD patients on chronic dopaminergic therapy. The consistent association of hallucinations with combined levodopa/agonist therapy suggests that these drugs may play a role in the pathophysiology of hallucinations.