ObjectivesOlfactory disorder is one of the sensory features that reflects a decline in cognitive function. However, olfactory changes and the discernibility of smell testing in the aging population have yet to be fully elucidated. Therefore, this study aimed to examine the effectiveness of the Chinese Smell Identification Test (CSIT) in distinguishing individuals with cognitive decline from those with normal aging and to determine whether the patients with MCI and AD show changes in their olfactory identification abilities.MethodsThis cross-sectional study included eligible participants aged over 50 years between October 2019 and December 2021. The participants were divided into three groups: individuals with mild cognitive impairment (MCI), individuals with Alzheimer's disease (AD), and cognitively normal controls (NCs). All participants were assessed using neuropsychiatric scales, the Activity of Daily Living scale, and the 16-odor cognitive state test (CSIT) test. The test scores and the severity of olfactory impairment were also recorded for each participant.ResultsIn total, 366 eligible participants were recruited, including 188 participants with MCI, 42 patients with AD, and 136 NCs. Patients with MCI achieved a mean CSIT score of 13.06 ± 2.05, while patients with AD achieved a mean score of 11.38 ± 3.25. These scores were significantly lower than those of the NC group (14.6 ± 1.57; P < 0.001). An analysis showed that 19.9% of NCs exhibited mild olfactory impairment, while 52.7% of patients with MCI and 69% of patients with AD exhibited mild to severe olfactory impairment. The CSIT score was positively correlated with the MoCA and MMSE scores. The CIST score and the severity of olfactory impairment were identified as robust indicators for MCI and AD, even after adjusting for age, gender, and level of education. Age and educational level were identified as two important confounding factors that influence cognitive function. However, no significant interactive effects were observed between these confounders and CIST scores in determining the risk of MCI. The area under the ROC curve (AUC) generated from the ROC analysis was 0.738 and 0.813 in distinguishing patients with MCI and patients with AD from NCs based on the CIST scores, respectively. The optimal cutoff for distinguishing MCI from NCs was 13, and for distinguishing AD from NCs was 11. The AUC for distinguishing AD from MCI was 0.62.ConclusionsThe olfactory identification function is frequently affected in patients with MCI and patients with AD. CSIT is a beneficial tool for the early screening of cognitive impairment among elderly patients with cognitive or memory issues.