Regional cortical thinning and cerebrospinal biomarkers predict worsening daily functioning across the Alzheimer's disease spectrum

Abstract

Background: Impairment in instrumental activities of daily living (IADL) heralds the transition from mild cognitive impairment (MCI) to dementia and is a major source of burden for both the patient and caregiver.

Objective: To investigate the relationship between IADL and regional cortical thinning and cerebrospinal fluid (CSF) Alzheimer's disease (AD) biomarkers cross-sectionally and longitudinally in clinically normal (CN) elderly, MCI, and mild AD dementia subjects.

Methods: Two hundred and twenty nine CN, 395 MCI, and 188 AD dementia subjects participating in the Alzheimer's Disease Neuroimaging Initiative underwent baseline magnetic resonance imaging, baseline lumbar puncture, and clinical assessments, including the Functional Activities Questionnaire used to measure IADL, every 6 to 12 months up to 3 years. General linear regression and mixed effects models were employed.

Results: IADL impairment was associated with the interactions between lower inferior temporal cortical thickness and diagnosis (p < 0.0001), greater lateral occipital cortical thickness and diagnosis (p < 0.0001), and greater amyloid-β 1-42 (Aβ1-42) and diagnosis (p = 0.0002) at baseline (driven by AD dementia). Lower baseline supramarginal (p = 0.02) and inferior temporal (p = 0.05) cortical thickness, lower Aβ1-42 (p < 0.0001), and greater total tau (t-tau) (p = 0.02) were associated with greater rate of IADL impairment over time.

Conclusions: Temporal atrophy is associated with IADL impairment in mild AD dementia at baseline, while baseline parietal and temporal atrophy, lower CSF Aβ1-42, and greater t-tau predict worsening IADL impairment over time across the AD spectrum. These results emphasize the importance of assessing IADL at the stage of MCI and even at the transition from CN to MCI.

Keywords: Alzheimer's disease; cerebrospinal fluid; instrumental activities of daily living; magnetic resonance imaging; mild cognitive impairment.

Figure 1

Values predicted from general linear model of baseline FAQ regressed on diagnostic group and inferior temporal cortical thickness. Lines indicate the predicted values for FAQ, and symbols denote corresponding actual values. For simplicity, the graph was not adjusted for the negligible effects of additional partialed significant predictors in the model. AD (Alzheimer's disease), CN (clinically normal elderly), FAQ (Functional Activities Questionnaire), MCI (mild cognitive impairment).

Figure 2

Predicted values from fixed effects of best fitting longitudinal model of FAQ across time in the study vs. supramarginal cortical thickness for selected FAQ baselines by diagnostic groups. In each panel, one set of lines begins at a baseline FAQ one standard deviation below the mean (approximately zero) and the other starts at one standard deviation above the baseline FAQ mean (approximately 12). Other predictors were set to their mean and sex was set to female. AD (Alzheimer's disease), CN (clinically normal elderly), FAQ (Functional Activities Questionnaire), MCI (mild cognitive impairment).

Figure 3

Predicted values from fixed effects of best fitting longitudinal model of FAQ across time in the study vs. Aβ_1-42 for selected FAQ baselines by diagnostic groups. In each panel, one set of lines begins at a baseline FAQ one standard deviation below the mean (approximately zero) and the other starts at one standard deviation above the baseline FAQ mean (approximately 12). Other predictors were set to their mean and sex was set to female. AD (Alzheimer's disease), CN (clinically normal elderly), FAQ (Functional Activities Questionnaire), MCI (mild cognitive impairment).