Association of the Informant-Reported Memory Decline With Cognitive and Brain Deterioration Through the Alzheimer Clinical Continuum

Abstract

Background and objectives: Studies are sparse regarding the association between the informant-reported subjective memory decline (informant report) and Alzheimer disease (AD) biomarkers. This study thus aimed at determining the clinical relevance of the informant report throughout the AD clinical continuum, by assessing its specific relationships with amyloid deposition, cognition, and neurodegeneration.

Methods: Participants from the Imagerie Multimodale de la maladie d'Alzheimer à un stade Précoce (IMAP+) primary cohort and the Alzheimer Disease Neuroimaging Initiative (ADNI) replication cohort were included; all underwent multimodal neuroimaging and neuropsychological assessments. Follow-up data of IMAP+ participants over up to 36 months were also used for longitudinal analyses. The informant report was measured with the Cognitive Difficulties Scale (IMAP+) and Everyday Cognition (ADNI). General linear models were used to assess the cross-sectional associations between the informant report and amyloid-PET, cognitive performances, and neurodegeneration (atrophy and hypometabolism) in Alzheimer signature areas; while longitudinal links were assessed in IMAP+ with linear mixed-effects models.

Results: A total of 110 IMAP+ participants were included, including 32 cognitively unimpaired older individuals (controls, age: 70.91 ± 6.57 years, female: 50%), 25 patients with subjective cognitive decline (SCD, 65.88 ± 6.64, 40%), 35 with mild cognitive impairment (MCI, 72.49 ± 7.5, 34%), and 18 with Alzheimer-type dementia (AD dementia, 68.17 ± 8.59, 28%). Seven hundred thirty-one ADNI participants were included, including 157 controls (74.21 ± 5.95, 55%), 84 with SCD (72.00 ± 5.41, 63%), 369 with MCI (71.84 ± 7.4, 44%), and 121 with AD dementia (74.29 ± 7.75, 40%). In IMAP+, a higher informant report strongly correlated to greater amyloid-PET, specifically in patients with MCI (β = 0.48, p = 0.003), and to lower cognitive performance in patients with SCD (global cognition, β = -0.41, p = 0.04) and MCI (memory, β = -0.37, p = 0.03). Findings in patients with MCI were replicated in the ADNI (amyloid-PET, β = 0.25, p < 0.001; memory, β = -0.22, p < 0.001) and extended to neurodegeneration in AD signature areas (β = -0.2, p < 0.001). Longitudinal analyses in IMAP+ showed links with global cognitive decline over time in patients with MCI (estimate -0.74, SE 0.26, p = 0.005) and SCD (estimate -0.36, SE 0.26, p = 0.02) where a higher baseline informant report also predicted an increased amyloid-PET over time (estimate 0.008, SE 0.003, p = 0.02).

Discussion: Altogether, our findings suggest that the informant report is particularly relevant in patients with MCI where it strongly relates to higher amyloid-PET, indicative of impairment due to AD.

Trial registration information: ClinicalTrials.gov Identifier: NCT01638949.

Figure 1. Relationships at Baseline Between the I-SMD and Amyloid Deposition

Graphs illustrate the results of general linear models for the cross-sectional correlations between the I-SMD w-scores and either the global Florbetapir SUVr across the different stages of the AD clinical continuum (A. p_unadjusted < 0.05 in purple) or the Florbetapir-PET in the group of patients with MCI (B. Thresholded at p < 0.001 with a cluster-level correction for multiple comparisons; k > 582 voxels; no significant association was found in the other groups). AD = Alzheimer disease; CDS = Cognitive Difficulties Scale Questionnaire; I-SMD = informant-reported subjective memory decline; MCI = mild cognitive impairment; SCD = subjective cognitive decline; SUVr = standardized uptake value ratio.

Figure 2. Relationships at Baseline Between the I-SMD and Objective Cognitive or Memory Performances Across the Different Stages of the AD Clinical Continuum

Graphs illustrate the results of general linear models for the correlations between I-SMD w-scores and either global cognition (A. MMSE w-scores) or verbal episodic memory (B. ESR free recall w-scores) at baseline. p_unadjusted < 0.10 in pink and p_unadjusted < 0.05 in purple. AD = Alzheimer disease; CDS = Cognitive Difficulties Scale; ESR = Encoding, Storage, and Retrieval; I-SMD = informant-reported subjective memory decline; MCI = mild cognitive impairment; MMSE = Mini-Mental State Examination; SCD = subjective cognitive decline.

Figure 3. Relationships at Baseline Between I-SMD and Amyloid Deposition and Objective Cognitive or Memory Performances Within the ADNI Replication Cohort

Graphs illustrate the results of general linear models for the cross-sectional correlations between the I-SMD w-scores and either the global Florbetapir SUVr across the different stages of the AD clinical continuum (A. p_unadjusted < 0.05 in purple), the Florbetapir-PET in the group of patients with MCI (B. Thresholded at p < 0.001 with a cluster-level correction for multiple comparisons (red) and at p_FWE < 0.05 with a k > 200 voxels (orange); no significant association was found in the other groups), global cognition (C. MMSE w-scores), or verbal episodic memory (D. RAVLT immediate recall w-scores) at baseline. p_unadjusted < 0.10 in pink and p_unadjusted < 0.05 in purple. AD = Alzheimer disease; ADNI = Alzheimer Disease Neuroimaging Initiative; CDS = Cognitive Difficulties Scale; ECog = Everyday Cognition Questionnaire; I-SMD = informant-reported subjective memory decline; MCI = mild cognitive impairment; MMSE = Mini-Mental State Examination; RAVLT = Rey Auditory Verbal Learning Test; SCD = subjective cognitive decline; SUVr = standardized uptake value ratio.

Figure 4. Relationships at Baseline Between I-SMD and Glucose Metabolism and Gray Matter Volume in Alzheimer Signature Areas Within the ADNI Replication Cohort

Graphs illustrate the results of general linear models for the cross-sectional correlations between the I-SMD w-scores and either the glucose metabolism (A) or gray matter volume (B) in Alzheimer signature areas determined in a previous study across the different stages of the AD clinical continuum (p_unadjusted < 0.05 in purple). Results of the whole-brain voxelwise correlations between the I-SMD score and glucose metabolism and gray matter volume are provided in eFigure 1 and detailed in eTables 3 (IMAP+) and 6 (ADNI) (links.lww.com/WNL/C777). AD = Alzheimer disease; ADNI = Alzheimer's Disease Neuroimaging Initiative; FDG = ¹⁸F-fluorodeoxyglucose; I-SMD = informant-reported subjective memory decline.