Résumé
<p><b>INTRODUCTION</b>Mild cognitive impairment (MCI) is an important clinical entity with significant management implications. However, traditional screening tools lack the sensitivity needed to detect amnestic MCI (MCI-A). Montreal Cognitive Assessment (MoCA) has yet to be validated for the diagnosis of MCI in a multiracial society such as Singapore. We thus aimed to study the effectiveness of MoCA for the diagnosis of MCI-A in the Singapore population.</p><p><b>METHODS</b>Data on patients with MCI-A and mild Alzheimer's disease (AD) was obtained from a prospectively collected clinical database between January 2008 and January 2011. Patients with no cognitive impairment (NCI) were recruited from among the spouses and friends of patients attending the memory clinic.</p><p><b>RESULTS</b>There were a total of 212 participants (103 NCI, 49 MCI-A, 60 mild AD). For the diagnosis of MCI-A, a MoCA score of < 26 for patients with ≤ 10 years of education, and a score of < 27 for patients with > 10 years of education provided a sensitivity of > 94%. For the diagnosis of mild AD, a MoCA score of < 24 for patients with ≤ 10 years of education, and a score of < 25 for patients with > 10 years of education provided a sensitivity of > 85%.</p><p><b>CONCLUSION</b>In the Singapore population, we recommend cutoff scores of 26/27 and 24/25 be used to detect MCI-A and mild AD, respectively, when using MoCA. For patients with ≤ 10 years of education, a +1 point correction is needed.</p>