RESUMO
OBJECTIVE: To determine the diagnostic utility of olfactory testing in patients with neurodegenerative parkinsonism. METHODS: The Sniffin' Sticks test battery for assessment of odor identification, discrimination, and threshold was applied to patients with Parkinson's disease (PD), multiple system atrophy (MSA) and progressive supranuclear palsy (PSP) as well as healthy controls (HC). Two different cohorts were analyzed: A PD/healthy control that included PD patients and HC as well as a PD/diseased control cohort for which patients PD, MSA and PSP were recruited. The former cohort was exploited to calculate cut-off values that discriminate PD patients from HC with a sensitivity (sensitivity-weighted cut-off) or specificity (specificity-weighted cut-off) exceeding 95%, respectively. The PD/diseased controls cohort was used to determine the diagnostic accuracy using these cut-off values in discriminating patients with neurodegenerative parkinsonism. RESULTS: PD patients (n = 67) performed significantly worse in olfactory testing than HC (n = 41) and patients with MSA (n = 23) or PSP (n = 23). There was no significant difference in olfactory function between MSA and PSP patients. Diagnostic performance of the identification subscore was similar to the sum score of the Sniffin' Sticks test (AUC identification test 0.94, AUC sum score 0.96), while threshold and discrimination subscores were inferior. In patients with parkinsonism, the specificity-weighted cut-off predicted a diagnosis of PD with a sensitivity and specificity of 76.6 and 87.0%, respectively. The discriminative value of this cut-off in separating PD from MSA was 76.7% (sensitivity) and 95.7% (specificity). The corresponding, prevalence-adjusted positive predictive value of olfactory testing exceeded 95%. CONCLUSIONS: Our data suggest that assessment of olfactory function, particularly odor identification, can be useful to discriminate PD from atypical parkinsonian disorders, particularly MSA patients.