Fax +41 61 306 12 34 E-Mail karger@karger.ch www.karger.com Original Research Article Dement Geriatr Cogn Disord 2009;27:173–181 DOI: 10.1159/000200465 Predictive Value of Mild Cognitive Impairment for Dementia The Influence of Case Definition and Age M.A.E. Baars a M.P.J. van Boxtel a J.B. Dijkstra a P.J. Visser a, c M. van den Akker b F.R.J. Verhey a J. Jolles a a Department of Psychiatry and Neuropsychology and b Department of General Practice, School for Public Health and Primary Care CAPHRI, Faculty of Health, Medicine and Life Sciences, Maastricht University, Maastricht, and c Department of Neurology, VU University Medical Centre, Amsterdam, The Netherlands Introduction Mild cognitive impairment (MCI) is a transitional state between normal functioning and dementia [1]. Pro- gression to dementia is one possible outcome for indi- viduals classified as having MCI. The rate of progression to dementia differs greatly between studies [2, 3]. Case definition is one reason for these differences. In the ear- lier studies, amnestic and nonamnestic MCI subtypes were predominantly used as predictors of Alzheimer’s disease [1, 4]. Over the past years, the discussion on the applicability of MCI as a prognostic marker has shifted towards a broader interpretation that includes several do- mains of cognition, such as memory and executive func- tioning in 1 definition [5]. The predictive value of MCI for dementia has been found to be highest when a com- bination of complaints (reported by the patient or part- ner), cognitive performance in multiple domains, and preserved overall general functioning are taken into ac- count [6]. However, it has also been suggested that subjec- tive complaints should not be included in the definition of MCI, because the predictive value for dementia does not increase when these complaints are taken into ac- count [7] . Key Words Dementia  Age  Cognition  Mild cognitive impairment  Population-based study Abstract Background/Aims: In population studies, different mild cognitive impairment (MCI) definitions have been used to predict dementia at a later stage. This study compared pre- dictive values of different MCI definitions for dementia, and the effect of age on the predictive values was investigated. Methods: This study was conducted as part of an ongoing longitudinal study into the determinants of cognitive aging, the Maastricht Aging Study. Results: MCI best predicted de- mentia when multiple cognitive domains were considered and subjective complaints were not (sensitivity: 0.66, speci- ficity: 0.78). Age had a strong influence on the sensitivity of MCI for dementia (age 60–70 years: sensitivity = 0.56; age 70–85 years: sensitivity = 0.70). Conclusion: The inclusion of multiple cognitive domains and participants aged 70 years and older leads to the best prediction of dementia, regard- less of subjective complaints. Copyright © 2009 S. Karger AG, Basel Accepted: November 12, 2008 Published online: February 7, 2009 M.A.E. Baars, MSc School for Mental Health and Neuroscience, Division of Brain and Cognition Maastricht University, DRT 12, PO Box 616 NL–6200 MD Maastricht (The Netherlands) Tel. +31 43 388 4088, Fax +31 43 388 4092, E-Mail lia.baars@np.unimaas.nl © 2009 S. Karger AG, Basel 1420–8008/09/0272–0173$26.00/0 Accessible online at: www.karger.com/dem Downloaded by: Vrije Universiteit 130.37.164.140 - 12/9/2014 10:48:43 AM