Comparing the accuracy of different smell identification tests in Parkinson's disease: relevance of cultural aspects.

OBJECTIVE

The aim of this study is to determine the usefulness of the University of Pennsylvania smell identification test (UPSIT), sniffin sticks (SS-16) and brief smell identification test (B-SIT) to assess smell identification in the Mexican population and its accuracy in discriminating subjects with Parkinson's disease (PD).

METHODS

We included 199 nondemented PD subjects and 199 control subjects matched by gender. Smell identification was tested using the UPSIT and SS-16. Our group obtained B-SIT data from a previous report.

RESULTS

The mean number of UPSIT items correctly identified by controls was 27.3±6; the PD group had a mean score of 19.4±6. UPSIT had a sensitivity of 82% with a specificity of 66% for a cut-off score of ≤25 for detection of PD. The mean number of SS-16 items correctly identified by controls was 10.3±2.2, while the PD group had 7.4±2.8 correct answers. For SS-16, sensitivity was 77.8% and specificity of 71.2% when using a cut-off value of ≤9. Lemon, turpentine and rose had an identification rate below the 25th percentile for all three tests. Odors with an identification rate above the 75th percentile include banana for all three tests, and gasoline, onion and chocolate for UPSIT and B-SIT.

CONCLUSION

The sensitivity and specificity of the smell tests that were evaluated were lower in comparison to other published reports. Cultural biases and smell familiarity may influence the test results. The development of a true cross-culturally adapted smell identification test is warranted may improve test accuracy.