Elsevier

The Lancet Neurology

Volume 15, Issue 10, September 2016, Pages 1044-1053
The Lancet Neurology

Articles
Clinical and cognitive trajectories in cognitively healthy elderly individuals with suspected non-Alzheimer's disease pathophysiology (SNAP) or Alzheimer's disease pathology: a longitudinal study

https://doi.org/10.1016/S1474-4422(16)30125-9Get rights and content

Summary

Background

Brain amyloid β (Aβ) deposition and neurodegeneration have been documented in about 50–60% of cognitively healthy elderly individuals (aged 60 years or older). The long-term cognitive consequences of the presence of Alzheimer's disease pathology and neurodegeneration, and whether they have an independent or synergistic effect on cognition, are unclear. We aimed to characterise the long-term clinical and cognitive trajectories of healthy elderly individuals using a two-marker (Alzheimer's disease pathology and neurodegeneration) imaging construct.

Methods

Between Nov 3, 2006, and Nov 25, 2014, 573 cognitively healthy individuals in Melbourne and Perth, Australia, (mean age 73·1 years [SD 6·2]; 58% women) were enrolled in the Australian Imaging, Biomarker and Lifestyle (AIBL) study. Alzheimer's disease pathology (A) was determined by measuring Aβ deposition by PET, and neurodegeneration (N) was established by measuring hippocampal volume using MRI. Individuals were categorised as AN, A+N, A+N+, or suspected non-Alzheimer's disease pathophysiology (AN+, SNAP). Clinical progression, hippocampal volume, standard neuropsychological tests, and domain-specific and global cognitive composite scores were assessed over 6 years of follow-up. Linear mixed effect models and a Cox proportional hazards model of survival were used to evaluate, compare, and contrast the clinical, cognitive, and volumetric trajectories of patients in the four AN categories.

Findings

50 (9%) healthy individuals were classified as A+N+, 87 (15%) as A+N, 310 (54%) as AN, and 126 (22%) as SNAP. APOE ε4 was more frequent in participants in the A+N+ (27; 54%) and A+N (42; 48%) groups than in the AN (66; 21%) and SNAP groups (23; 18%). The A+N and A+N+ groups had significantly faster cognitive decline than the AN group (0·08 SD per year for AIBL-Preclinical AD Cognitive Composite [PACC]; p<0·0001; and 0·25; p<0·0001; respectively). The A +N+ group also had faster hippocampal atrophy than the AN group (0·04 cm3 per year; p=0·02). The SNAP group generally did not show significant decline over time compared with the AN group (0·03 SD per year [p=0·19] for AIBL-PACC and a 0·02 cm3 per year increase [p=0·16] for hippocampal volume), although SNAP was sometimes associated with lower baseline cognitive scores (0·20 SD less than AN for AIBL-PACC). Within the follow-up, 24% (n=12) of individuals in the A+N+ group and 16% (n=14) in the A+N group progressed to amnestic mild cognitive impairment or Alzheimer's disease, compared with 9% (n=11) in the SNAP group.

Interpretation

Brain amyloidosis, a surrogate marker of Alzheimer's disease pathology, is a risk factor for cognitive decline and for progression from preclinical stages to symptomatic stages of the disease, with neurodegeneration acting as a compounding factor. However, neurodegeneration alone does not confer a significantly different risk of cognitive decline from that in the group with neither brain amyloidosis or neurodegeneration.

Funding

CSIRO Flagship Collaboration Fund and the Science and Industry Endowment Fund (SIEF), National Health and Medical Research Council, the Dementia Collaborative Research Centres programme, McCusker Alzheimer's Research Foundation, and Operational Infrastructure Support from the Government of Victoria.

Introduction

The National Institute on Aging–Alzheimer's Association's (NIA-AA) criteria for preclinical Alzheimer's disease1 classify individuals into one of three stages based on two types of markers—those specific for amyloid β (Aβ; either low Aβ1−42 levels in CSF or high Aβ levels in the brain as measured by PET), and those for neuronal injury (increased total tau in CSF; Alzheimer's disease-like glucose hypometabolism as assessed by 18F-fluorodeoxyglucose [FDG] PET or brain atrophy as measured by structural MRI) and grouped as neurodegeneration. Based on these markers, stage 1 is defined as the presence of Alzheimer's disease pathology (namely brain Aβ deposition); stage 2 as the presence of Alzheimer's disease pathology plus neurodegeneration; and stage 3 as the presence of Alzheimer's disease pathology and neurodegeneration accompanied by subtle deficits in cognition.1, 2, 3

Application of the NIA-AA criteria to the healthy elderly individuals enrolled in the Mayo Clinic Study of Aging (MCSA)4 showed that about 66% of 450 participants did not fit any of the three categories. 43% of 450 participants had no positive marker of Alzheimer's disease pathology or neurodegeneration, 23% were classified with neurodegeneration without evidence of Alzheimer's disease pathology, 16% had evidence of Alzheimer's disease pathology (stage 1), 12% had Alzheimer's disease pathology and neurodegeneration (stage 2), 3% with Alzheimer's disease pathology and neurodegeneration also showed subtle cognitive impairment, while 3% remained unclassified. On the basis of these observations, Jack and colleagues4 proposed the addition of two new categories to the preclinical criteria: stage 0, comprising cognitively healthy elderly individuals with no evidence of Alzheimer's disease pathology or neurodegeneration, and a group termed suspected non-Alzheimer's disease pathophysiology (SNAP) consisting of elderly adults with Alzheimer's disease-like neurodegeneration but no evidence of Alzheimer's disease pathology.

Research in context

Evidence before this study

We searched PubMed and Google Scholar, with the terms “Preclinical AD criteria”, “SNAP”, and “non-amyloid pathology” to identify studies published in English between Jan 1, 2011, and April 1, 2016. Cognitively healthy cohorts have been examined using the two biomarker construct definitions of preclinical Alzheimer's disease (AN [no Alzheimer's disease pathology or neurodegeneration], SNAP [suspected non-Alzheimer's disease pathophysiology], A+N [evidence of Alzheimer's disease pathology but no evidence of neurodegeneration], and A+N+ [Alzheimer's disease pathology and neurodegeneration]). The prevalence of each of these categories was consistent across the different studies. Despite differences in the design, cross-sectional or longitudinal, or in the length of the follow-up, these studies were also consistent in findings, across multiple cognitive tests and domains, in which the A+N+ group had the fastest decline, followed by the A+N group, and in which the SNAP group (although exhibiting worse baseline cognitive performance) had similar rates of decline to the AN group.

Added value of this study

To the best of our knowledge, this study has the longest follow-up assessing the cognitive trajectories in preclinical Alzheimer's disease using a two imaging marker construct. This longer assessment allowed a better understanding of disease progression, changes in the two marker classification, and the effect on the cognitive and volumetric trajectories of individuals at the peri-thresholds for the two-biomarker constructs. Also, the fairly large sample size has enabled assessment of the effect of APOE ε4 status on the two-biomarker construct. This study confirms that SNAP constitutes a separate pathophysiological entity or group of disorders with a cognitive and clinical trajectory different from the Alzheimer's disease pathway.

Implications of all of the available evidence

Brain amyloidosis, a surrogate marker of Alzheimer's disease pathology, is a risk factor for cognitive decline and for progression from preclinical stages to symptomatic stages of Alzheimer's disease, with neurodegeneration acting as a compounding factor. However, neurodegeneration alone does not confer a significantly different risk from that in the AN control group, and individuals with neurodegeneration but not detectable brain amyloidosis are not on the Alzheimer's disease pathway.

The data from the MCSA study4 suggest that the risk for Alzheimer's disease in cognitively normal elderly adults can be staged based on a two-biomarker construct; individuals with no Alzheimer's disease pathology or neurodegeneration (AN), evidence of Alzheimer's disease pathology but no evidence of neurodegeneration (A+N), neurodegeneration but no Alzheimer's disease pathology (AN+ or SNAP), and both Alzheimer's disease pathology and neurodegeneration (A+N+). Several cognitively healthy cohorts have been examined using these constructs,5, 6, 7, 8 and although the definitions were originally provided for preclinical Alzheimer's disease, the examinations have been extended to individuals with mild cognitive impairment.9, 10, 11 There was a remarkable consistency in the prevalence of these categories across the different studies, which shows the robustness of this two-construct classification, especially in view of the differences between cohorts, sampling, thresholds, and analytical approaches.4, 5, 6, 7, 8, 9, 10, 11

Jack and colleagues12 recently applied this two marker classification approach to the general population and noted that, by age 50 years, almost none of the participants had detectable Alzheimer's disease pathology or neurodegeneration, whereas by age 89 years, about 80% of the study population had Alzheimer's disease pathology, neurodegeneration, or both. Further, three of the studies assessing cognitively healthy cohorts also assessed the effect of the two-biomarker construct on longitudinal cognitive performance.6, 7, 8 These studies reported that longitudinal performance of the SNAP group did not differ significantly from that of the AN group.

Against this background, we were interested in assessing the long-term (6 year) clinical and cognitive trajectories of the four groups, to try to elucidate the independent, interactive, or additive effects of Alzheimer's disease pathology and neurodegeneration on disease progression in participants of the Australian Imaging, Biomarker and Lifestyle (AIBL) study3 who were cognitively healthy at baseline examination and assigned to one of the AN categories based on Aβ imaging with PET and hippocampal volume with MRI. The longitudinal trajectories of different cognitive domains across the four categories, as well as the effects of APOE ε4 status, were explored. Further, over the 6 years of follow-up, we assessed the progression through the A and N categories and the risk of progression to mild cognitive impairment or Alzheimer's disease.

Section snippets

Study design and participants

The AIBL study is a cohort study of ageing, integrating data from neuroimaging, biomarkers, lifestyle, clinical, and neuropsychological assessments. Participants were recruited by two study centres in Melbourne, VIC, and Perth, WA, Australia. Participants with mild cognitive impairment and Alzheimer's disease were recruited from tertiary Memory Disorders Clinics or from primary-care physicians, cognitively healthy individuals (healthy controls) were recruited through advertisement or from

Results

At baseline, the mean age of the 573 AIBL study participants was 73·1 years (SD 6·2) and 330 (58%) were women. Based on the two-imaging construct model, 310 (54%) of 573 healthy controls from AIBL were assigned to the AN (control) category, 126 (22%) to the SNAP category, 87 (15%) to the A+N category, and 50 (9%) to the A+N+ category (table).

Age, years of education, APOE ε4 status, and disease progression were significantly different between the A+N and the AN groups, and between the A+N+

Discussion

The assessment of the two-biomarker construct in a large cohort over 6 years of follow-up showed that individuals classified as SNAP have qualitatively and quantitatively different clinical, cognitive, and volumetric trajectories than do those with Alzheimer's disease pathology. Moreover, this study provides further evidence that by using a simple two-imaging marker cross-sectional classification, it is possible to differentiate elderly individuals on the Alzheimer's disease pathway from those

References (33)

  • JL Bernal-Rusiel et al.

    Statistical analysis of longitudinal neuroimage data with linear mixed effects models

    Neuroimage

    (2013)
  • YY Lim et al.

    APOE ɛ4 moderates amyloid-related memory decline in preclinical Alzheimer's disease

    Neurobiol Aging

    (2015)
  • CR Jack et al.

    An operational approach to National Institute on Aging–Alzheimer's Association criteria for preclinical Alzheimer disease

    Ann Neurol

    (2012)
  • D Knopman et al.

    Short-term clinical outcomes for stages of NIA-AA preclinical Alzheimer disease

    Neurology

    (2012)
  • EC Mormino et al.

    Synergistic effect of β-amyloid and neurodegeneration on cognitive decline in clinically normal individuals

    JAMA Neurol

    (2014)
  • AC van Harten et al.

    Preclinical AD predicts decline in memory and executive functions in subjective complaints

    Neurology

    (2013)
  • Cited by (0)

    View full text