Odorant Item Specific Olfactory Identification Deficit May Differentiate Alzheimer Disease From Aging

•Screening for amnestic disorder must differentiate changes associated with normal aging.•Using an odorant identification test, the authors identified specific odorants which best differentiate amnestic disorder from normal aging. The identified odorants were validated with performance on neuropsychological testing and prediction of conversion from amnestic mild cognitive impairment to Alzheimer disease.•Olfactory identification impairment was associated with poor performance in numerous cognitive domains and was predictive of progression of mild cognitive impairment.•Olfactory identification deficit can be utilized for risk stratification of progression from amnestic mild cognitive impairment (aMCI) to Alzheimer disease (AD) and to enrich clinical trials of potential disease-modifying therapies for converters. To explore whether the ability to recognize specific odorant items is differentially affected in aging versus Alzheimer disease (AD); to refine olfactory identification deficit (OID) as a biomarker of prodromal and early AD. Prospective multicenter cross-sectional study with a longitudinal arm. Outpatient memory diagnostic clinics in New York and Texas. Adults aged 65 and older with amnestic mild cognitive impairment (aMCI) and AD and healthy aging (HA) subjects in the comparison group. Participants completed the University of Pennsylvania Smell Identification Test (UPSIT) and neuropsychological testing. AD-associated odorants (AD-10) were selected based on a model of ordinal logistic regression. Age-associated odorants (Age-10) were identified using a linear model. For the 841 participants (234 HA, 192 aMCI, 415 AD), AD-10 was superior to Age-10 in separating HA and AD. AD-10 was associated with a more widespread cognitive deficit across multiple domains, in contrast to Age-10. The disease- and age-associated odorants clustered separately in age and AD. AD-10 predicted conversion from aMCI to AD. Nonoverlapping UPSIT items were identified that were individually associated with age and disease. Despite a modest predictive value of the AD-specific items for conversion to AD, the AD-specific items may be useful in enriching samples to better identify those at risk for AD. Further studies are needed with monomolecular and unilateral stimulation and orthogonal biomarker validation to further refine disease- and age-associated signals.