Premium
P3–137: Diet and cognition: Baseline associations in the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER)
Author(s) -
Lehtisalo Jenni,
Lindström Jaana,
Ngandu Tiia,
Laatikainen Tiina,
Soininen Hilkka,
Strandberg Timo,
Tuomilehto Jaakko,
Kivipelto Miia
Publication year - 2013
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2013.05.1208
Subject(s) - medicine , micronutrient , cognition , dementia , gerontology , effects of sleep deprivation on cognitive performance , physical therapy , disease , pathology , psychiatry
Background: Recent studies highlight the utilization of biomarkers for predicting the onset of symptomatic Alzheimer disease (AD) in cognitively normal individuals. Longitudinal analyses have demonstrated increased rates of hippocampal volume loss with progression to symptomatic AD. This study compared molecular and volumetric biomarkers (Ab 42, hippocampal volume) to predict time to incident cognitive impairment in a cognitively normal sample.Methods: Longitudinal data from cognitively normal participants studied in the Knight Alzheimer Disease Research Center for a mean of 3.70 years (SD1⁄41.46). Participants had amyloid imaging and CSF collection within one year of a clinical assessment indicating normal cognition. Imaging data was processed using in-house novel software and the FreeSurfer (FS) software suite. FS regions of interest were applied to volumetric T1-weighted MRI images and PIB data for mean-cortical binding potential (MCBP). Hippocampal volume (HCV), adjusted for head size, was classified as abnormal if greater than one standard deviation from the age-dependent mean. Additional cutoffs for abnormal biomarker levels were established (PIB positive MCBP>0.18, CSF+ AB42<500). Cox proportional hazards models tested whether individual biomarkers were related to time to incident cognitive impairment when adjusting for age, gender, education, and race. Results: Of the 176 participants who met inclusion criteria, 22 developed cognitive impairment over the study period. In separate models, time to cognitive impairment was marginally associated with MCBP (p1⁄4.089) and normalized hippocampal volume (nHCV) (p1⁄4.003). The model containing both MCBP and nHCV attenuated the effect of MCBP but not nHCV (p1⁄4.340 and .004, respectively). There was no interaction effect between MCBP and nHCV (p1⁄4.238). The nHCV model showed better predictive accuracy (CPE1⁄4.798, SE1⁄4.032) than the model containing MCBP (CPE1⁄4.778, SE1⁄4.035). Similar predictive accuracy was found for the model containing both MCBP and nHCV (CPE1⁄4.797, SE1⁄4.032). Conclusions: Abnormally low nHCV values reflect early changes in the spectrum of AD. While biomarkers alone likely will not provide individual prediction information for development of symptomatic AD, they are a valuable tool in assessing risk for developing clinical symptoms.