Â鶹´«Ã½Ó³»­

Â鶹´«Ã½Ó³»­

Explore the latest content from across our publications
Join The Â鶹´«Ã½Ó³»­

Log In

Forgot Password?
Create New Account

Loading... please wait

Abstract Details

Role of University of Pennsylvania Smell Identification Test (UPSIT) in Diagnosis of Parkinsonian Syndromes in a Real World Setting
Movement Disorders
P8 - Poster Session 8 (8:00 AM-9:00 AM)
3-007
To determine the utility of the University of Pennsylvania Smell Identification Test (UPSIT) in predicting Parkinson’s disease (PD) versus non-Parkinson’s disease (non-PD) final diagnosis by a movement disorder specialist
UPSIT is a "scratch and sniff" smell test comprised of 40 smells designed to detect hyposmia. Severe olfactory dysfunction is a common early symptom of PD but is not characteristic of other parkinsonian syndromes, rendering UPSIT a potentially useful tool in the early differentiation of PD from non-PD syndromes. 
We created a patient database including 81 movement disorder patients who have completed the UPSIT at our academic medical center. Employing logistic regression analyses, we created three models, each with a different combination of independent variables, to predict final diagnosis of PD versus non-PD by a movement disorder specialist. Our first model utilized UPSIT percentile as the sole independent variable. Our second model used initial diagnosis (defined as diagnosis by a movement disorder specialist prior to UPSIT administration and levodopa challenge). Our third model included a combination of both UPSIT percentile and initial diagnosis. We conducted chi-squared goodness-of-fit analyses, evaluated the contributions of each independent variable to each model’s output, and performed a ROC (receiver operating characteristic) analysis for each model.
The UPSIT percentile model, the initial diagnosis model, and the combined model were all statistically significant by the chi-squared goodness-of-fit-analyses (p=0.004550365, 9.474786e-07, and 5.907252e-08, respectively). Each independent variable significantly contributed in each of our models. This illustrates all three models were capable of accurately predicting the final diagnosis. The Area Under the Curve determined by our ROC analyses for the first, second, and third models were 0.718, 0.754, and 0.868, respectively. This suggests that UPSIT percentile significantly improves prediction of final diagnosis.
UPSIT percentile is a useful, convenient, and low-cost clinical tool for early differentiation of PD from other parkinsonism syndromes.
Authors/Disclosures
Nicole Pulley, MD (John’s Hopkins Hospital)
PRESENTER 		Dr. Pulley has nothing to disclose.
No disclosure on file
No disclosure on file
Mustafa S. Siddiqui, MD, FÂ鶹´«Ã½Ó³»­ Dr. Siddiqui has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Boston Scientific Neuromodulation. Dr. Siddiqui has received personal compensation in the range of $500-$4,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Medtronic. The institution of Dr. Siddiqui has received research support from Boston Scientific Neuromodulation. The institution of Dr. Siddiqui has received research support from Abbvie. The institution of Dr. Siddiqui has received research support from National Institute of Health .