Olfactory Dysfunction as a Window Into the Heterogeneity of Parkinson Disease

Abstract

Background: While olfactory dysfunction is common in Parkinson disease (PD), its neural basis and clinical implications remain to be clarified. We investigated the neural substrates and clinical profiles, particularly non-motor symptoms (NMSs), associated with olfactory function.

Methods: This retrospective study included 259 drug-naïve patients with PD who underwent the Cross-Cultural Smell Identification Test (CC-SIT), comprehensive autonomic function test, neuropsychological assessments, and the Neuropsychiatric Inventory Questionnaire (NPI-Q) at diagnosis. NMS profiles were compared across olfactory groups defined by CC-SIT scores (normosmia [n = 45], hyposmia [n = 143], anosmia [n = 74]). Associations between olfaction and clinical/imaging variables were assessed using correlation and path analyses. Cox proportional hazards models were employed to evaluate the risk of developing motor complications or PD dementia according to olfactory status.

Results: CC-SIT scores correlated with Composite Autonomic Severity Scale scores (rho = -0.219, p = 0.001), NPI-Q scores (rho = -0.269, p < 0.001), and cognitive performance in memory (rho = 0.288, p < 0.001) and frontal/executive domains (rho = 0.205, p = 0.001). Dopaminergic depletion in the caudate nucleus and limbic atrophy emerged as neural substrates underlying olfactory dysfunction, mediating the association between olfaction and cognitive/neuropsychiatric symptoms. Anosmia was associated with increased risk of developing PD dementia compared to normosmia (hazard ratio [HR]: 2.579; 95% confidence intervals [CI]: 1.137-5.851) and hyposmia (HR: 2.783; 95% CI: 1.437-5.390). Anosmia was associated with higher risk of developing freezing of gait (HR: 2.571; 95% CI: 1.077-6.134) compared to normosmia.

Conclusions: Olfactory dysfunction serves as a multifaceted clinical marker associated with convergent degeneration of nigrostriatal and limbic pathways, offering insights into PD phenotypic variance and its prognostic implications.