Smell Identification Deficits as a Predictive Tool for Schizophrenia

Abstract

Olfactory deficits in schizophrenia are well-established and related to the deficit syndrome (DS), or the presence of primary negative symptoms. We explored the predictive ability of smell identification deficits in identifying DS patients with diminished emotional expression (EE). The full 40-item University of Pennsylvania Smell Identification Test (UPSIT) was administered to 54 healthy controls and 52 patients with schizophrenia or schizoaffective disorder. A Positive and Negative Syndrome Scale (PANSS)-derived proxy measure was used to diagnose 13 DS patients. A Kruskal-Wallis Test revealed a significant difference in UPSIT scores among all groups, chi&sup2;(2, n = 106) = 30.19, p < .001. Planned comparisons using Mann-Whitney U tests revealed lower UPSIT total scores by DS patients vs. controls (U = 71.00, z = -4.46, p < .001, r = .54), non-DS patients vs. controls (U = 525.00, z = -4.13, p < .001, r = .43) and DS patients vs. non-DS patients (U = 142.00, z = -2.36, p = .018, r = .33). Receiver-operating-characteristic (ROC) curve analysis of UPSIT scores yielded 89.9% accuracy for identifying DS vs. controls (95% CI: 78.4 -- 1.01%). A criterion set to 29 or less for classification of DS obtained sensitivity of 84.6% and specificity of 92.6%. The UPSIT showed high accuracy for identifying DS patients with diminished EE, whose more severe SIDs implicate common underlying prefrontal and parietal regions involving facial expression and odor identification. Further investigation and development of measures to better identify early sensory deficits will promote early screening in schizophrenia.