Visual hallucinations in the diagnosis of idiopathic Parkinson's disease: a retrospective autopsy study

Background For many years, visual hallucinations (VH) in idiopathic Parkinson's disease (PD) were thought to be a complication of antiparkinsonian treatment. The cause of VH is now thought to be nerve-cell loss and Lewy-body pathology in the ventral-temporal regions of the brain. However, the use of VH as a clinical sign of PD has not been investigated. Methods 788 cases with parkinsonism archived at the Queen Square Brain Bank for Neurological Diseases were identified for study. Patients had not been given standardised antemortem assessments. Clinical records were assessed for reports of VH. The incidence of VH in pathologically diagnosed cases was calculated, and factors affecting onset of VH were investigated. Findings VH occurred in 50% (221/445) of patients with PD, in 73% (32/44) with dementia with Lewy bodies (DLB), and in only 7% (18/255) of patients with non-Lewy-body parkinsonism. The specificity of VH for Lewy-body parkinsonism (PD and DLB combined) was 92.9% (95% Cl 89.1-95-8) and the positive predictive value was 93.4% (89.7-95.8). VH were associated with cognitive dysfunction (hazard ratio 5.62, 3.37-9.35), autonomic dysfunction (3.13, 1.77-5.52), axial rigidity (2.22, 1.26-3.85) within the first 2 years of disease onset and also age of onset (1.05, 1.03-1.07). In PD, the onset of VH typically occurred in the second half of the disease course, and time to onset of VH was only weakly correlated with use of selegiline (Spearman's rho 0.22, p= 0.005) and ergot dopamine agonists (0.24, p=0.006) but not correlated with use of levodopa, amantadine, or anticholinergic drugs. Interpretation The presence of VH is helpful in the differentiation of PD from other non-Lewy-body causes of parkinsonism. We propose that VH be added, as a supportive criterion, to the operational clinical criteria for the diagnosis of PD.