Identification of odors, faces, cities and naming of objects in patients with subjective cognitive decline, mild cognitive impairment and Alzheimer´s disease: a longitudinal study.

ABSTRACTObjective:Recent studies have tried to find a reliable way of predicting the development of Alzheimer´s Disease (AD) among patients with mild cognitive impairment (MCI), often focusing on olfactory dysfunction or semantic memory. Our study aimed to validate these findings while also comparing the predictive accuracy of olfactory and semantic assessments for this purpose.

METHOD: Six hundred fifty patients (median age 68, 58% females) including controls, SCD (subjective cognitive decline), non-amnestic MCI (naMCI), amnestic MCI (aMCI), and AD patients were tested for olfactory dysfunction by means of odor identification testing and semantic memory. Of those 650 patients, 120 participants with SCD, naMCI, or aMCI at baseline underwent a follow-up examination after two years on average. Of these 120 patients, 12% had developed AD at follow-up (converters), while 88% did not develop AD at follow-up (non-converters).

RESULTS: Analysis showed a significant difference only for initial olfactory identification between converters and non-converters. Sensitivity of impairment of olfactory identification for AD prediction was low at 46.2%, although specificity was high at 81.9%. Semantic memory impairment at baseline was not significantly related to AD conversion, although, when naming objects, significant differences were found between AD patients and all other groups and between naMCI and aMCI patients compared to controls and SCD patients.

CONCLUSIONS: Objective olfactory assessments are promising instruments for predicting the conversion to AD among MCI patients. However, due to their low sensitivity and high specificity, a combination with other neuropsychological tests might lead to an improved predictive accuracy. Further longitudinal studies with more participants are required to investigate the usefulness of semantic memory tests in this case.