Objective: To compare the screening performance of existing abbreviated (4-12 item) versions of the 40-item University of Pennsylvania Smell Identification Test (UPSIT), and to identify a novel subset of UPSIT items with the greatest sensitivity for detecting hyposmia, out of 23,231,378 possible combinations.

Background: Reduced ability to detect and recognise smells (hyposmia) develops with normal ageing and features in several neurodegenerative diseases, including Parkinson’s disease (PD). The 40-item UPSIT is a widely used screening tool for detecting hyposmia, but is time-consuming and expensive when used on a large scale.

Methods: We retrospectively assessed abbreviated UPSIT test versions for their ability to detect hyposmia using data from 891 healthy participants from the PREDICT-PD cohort, who had completed the full UPSIT test. Hyposmia was defined as the lowest 15th centile of UPSIT scores according to age and gender. Abbreviated tests included Versions A and B of the 4-item Pocket Smell Test (PST) and the 12-item Brief Smell Identification Test (B-SIT). We then evaluated all 23,231,378 possible combinations of smaller smell subsets of 1-7 items from 40 total UPSIT items. Screening performances of ‘winning’ subsets were reassessed in an independent cohort of 191 PREDICT-PD participants. Finally, ‘winning’ smells were compared with those most commonly misidentified by 44 persons with PD compared to 891 PREDICT-PD controls.

Results: The B-SIT achieved a sensitivity/specificity of 83.1%/79.6%, whilst PST Versions A and B had a sensitivity/specificity of 76.8%/64.9% and 88.0%/45.5% respectively in detecting hyposmia. Our data-driven approach identified a set of 7 smells which matched/surpassed B-SIT performance across all parameters. Alternatively, a specific set of just 4 smells (Menthol, Clove, Gingerbread and Orange) possessed a superior screening performance (maximising for sensitivity and NPV) for hyposmia detection in both PREDICT-PD cohorts and PD participants when compared with the PST and B-SIT. The same 4 smells also featured amongst those most commonly misidentified by PD participants compared to healthy controls.

Conclusions: The use of specific smell subsets identified within our study may provide more accurate and cost-effective means for detecting hyposmia in large cohorts, allowing more targeted administration of the full UPSIT. An earlier version of this abstract has been accepted for the forthcoming ABN Annual Meeting, 9-11 May 2018. An earlier version of this abstract has been accepted for the forthcoming ABN Annual Meeting, 9-11 May 2018

« Back to 2018 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/screening-performance-of-abbreviated-versions-of-the-upsit-smell-test/