Autori: Simona Sacuiu, Magnus Sjögren, Boo Johansson, Deborah Gustafson, Ingmar Skoog

Editorial: Neurology, 65, p.1894-1900, 2005.

Rezumat:

Objective: To evaluate the utility of assessing four cognitive domains obtained from four information sources to identify individuals at risk for developing dementia, Alzheimer disease (AD), and vascular dementia (VaD). Methods: A representative population sample of nondemented 85-year-olds (n=313) from Gothenburg, Sweden, was examined regarding memory, language, and visuospatial and executive functions, using self- and key informant reports and neuropsychiatric and psychometric examinations. The sample was followed 3 years for incident dementia, AD, or possible
VaD.
Results: All cases of dementia, AD, and VaD were preceded by low performance in most domains irrespective of information source. Isolated low memory performance or low cognitive performance with unimpaired memory did not predict dementia. Positive predictive values increased with number of domains affected. Self- and key informant reports were less useful for predicting dementia than neuropsychiatric and psychometric examinations. The best positive predictive value (88%) was for low cognitive performance in all domains using neuropsychiatric and psychometric examinations;
however, sensitivity was only 18%. Conclusions: Although memory impairment was necessary to predict dementia, it was not sufficient. Other cognitive domains needed to be affected. Relying on self-reports or key informants for early detection
of dementia excluded a large group at risk. In addition, vascular dementia appeared to have a prodromal stage.

Cuvinte cheie: procese cognitive, dementa preclinica, studiu longitudinal // cognition, preclinical dementia, longitudinal study

URL: http://www.neurology.org/cgi/content/full/65/12/1894