BACKGROUND: Mild cognitive impairment (MCI) and Alzheimer's disease (AD) are screened to distinguish whether the cognitive decline in older adults is attributed to pathological causes rather than normal aging. OBJECTIVE: The purpose of this review was to analyze the diagnostic performance of the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) in screening for MCI and AD. METHODS: Electronic searches were performed on MEDLINE, EMBASE, CINAHL, and PsycArticles databases using the following keywords: dementia and ADAS-Cog. The Quality Assessment of Diagnostic Accuracy Studies-2 was used to check the risk of bias in the diagnostic studies. RESULTS: We reviewed 14 studies, including 3,875 patients who met the selection criteria. In 2,624 MCI patients from nine studies, the pooled sensitivity of ADAS-Cog was 0.80 (95% confidence interval [CI], 0.68-0.88), the pooled specificity was 0.84 (95% CI, 0.75-0.90), and the area under the curve of summary receiver-operating characteristic curves (SROC AUC) was 0.89 (SE = 0.03). In 2,517 AD patients from 10 studies, the pooled sensitivity and pooled specificity were 0.91 (95% CI, 0.86-0.95) and 0.93 (95% CI, 0.88-0.95), respectively, and the sROC AUC was 0.97 (SE = 0.01). Although sub-analyzed according to age and years of education, there was no significant difference in the predictive validity of the ADAS-Cog. CONCLUSION: The ADAS-Cog has high predictive validity as a screening tool in both MCI and AD and has better diagnostic performance in patients with AD. When early screening for AD is desired, ADAS-Cog is a first-stage screening tool that can be initially employed.