While many studies have tracked the mental abilities of healthy people who test positive for Alzheimer’s biomarkers, few have examined the very old over the long term. Now, a group led by Beth Snitz at the University of Pittsburgh details how Aβ accumulation and hippocampal atrophy correlate with cognitive performance over an average of 12 years in people with an average age of 86. They report that while amyloid-positive individuals were more likely to develop deficits in many cognitive domains, those who lost only hippocampal volume tended to experience just memory loss. The findings, published in the November 6 JAMA Neurology, dissect how different pathologies affect different cognitive abilities, and shed light on the long-term consequences of Suspected Non-Alzheimer Pathophysiology (SNAP).

  • AD-related biomarkers in very old, healthy people foreshadow long-term cognitive trajectories.
  • As in younger cohorts, Aβ accumulation predicts global cognitive decline.
  • Hippocampal atrophy without Aβ buildup, SNAP, correlates only with memory loss.

 “This study is unique in examining a very old cohort with long follow-up,” said David Knopman, Mayo Clinic, Rochester, Minnesota.

Cognitively healthy people with amyloidosis and some sign of neurodegeneration, be it high levels of tau in cerebrospinal fluid, brain hypometabolism, and/or hippocampal atrophy (Aβ+/ND+), run a higher risk of becoming cognitively impaired late in life than people with no pathology (Aβ–/ND–). Aβ+/ND– individuals have slightly less risk than Aβ+/ND+ people, but the fate of people who have only the neurodegenerative biomarker (Aβ–/ND+) are less clear (Jack et al., 2016). Several dedicated SNAP studies that measured hippocampal volumes, cortical glucose metabolism, or tau in the cerebrospinal fluid, however, found no effects of SNAP on future cognition (Sep 2015 newsBurnham et al., 2016Mormino et al., 2016; Soldan et al., 2016). 

Memory Decline.

Verbal memory test scores declined faster in Aβ-positive than Aβ-negative individuals. [Courtesy of Zhao et al., © 2017 American Medical Association. All rights reserved.]

To assess the fate of older people who tested positive for either amyloidosis, neurodegeneration, or both, first author Yujing Zhao analyzed data from the Gingko Evaluation of Memory Study (GEMS). GEMS enrolled 3,069 participants age 75 or older who had normal cognition or mild cognitive impairment (MCI). Zhao used data from the 175 volunteers who took part in an imaging substudy. They had an average age of 78. Of these, 140 were cognitively healthy at time of enrollment while 35 had MCI. At baseline, and annually thereafter, all underwent cognitive testing, while seven to nine years after enrollment, when they were 86 years old on average, they had amyloid PET and structural MRI scans. The researchers classified individuals as Aβ+ if their global cortical PiB uptake was 1.57 times or more of that in a cerebellar reference region, and as ND+ if they had hippocampal atrophy. 

At the time of imaging, 42 participants were classified as Aβ–/ND–, 32 were Aβ+/ND–, 35 deemed Aβ–/ND+, and 66 were Aβ+/ND+, a distribution similar to that of another cohort in their mid- to late-80s (Jack et al., 2014). Each class followed a distinct trajectory of cognitive decline (image above). Consistent with previous studies, Aβ+/ND+ individuals deteriorated the fastest across all cognitive domains tested, including memory, attention, reaction speed and vigilance, task-switching, reasoning, and verbal fluency, while the Aβ+/ND– group developed trouble with memory, attention, task-switching, and verbal fluency.

Notably, people in the Aβ–/ND+ (SNAP) group showed signs of memory decline. Their visual memory test scores deteriorated faster than did those of Aβ–/ND– individuals. “We saw greater separation among the biomarker groups when we did the analysis by adding a quadratic term,” noted Snitz. Previous studies had used linear fits for the data. “We know from longitudinal studies that cognitive change often doesn’t occur linearly, however, the follow-up needs to be long enough and the assessments frequent enough to test for patterns of acceleration or deceleration over time,” Snitz added.

“It is possible that their quadratic model captures decline better,” agreed Samantha Burnham from the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Australia. Burnham found that people classified as SNAP in the Australian Imaging, Biomarker, and Lifestyle (AIBL) study sometimes had slightly poorer composite cognitive scores than Aβ–/ND– individuals at baseline, but maintained those scores across the next seven years. Still, she was excited the new results mostly mirrored findings that Aβ+/ND– and Aβ+/ND+ individuals declined faster than Aβ–/ND–. Claudia Kawas, a GEM study investigator at the University of California, Irvine, emphasized the importance of age. She wondered if the older GEM cohort, 86 at time of imaging versus 73 in AIBL, might have different etiologies underlying SNAP, since various non-AD pathologies become more frequent and worsen with age.

Snitz’s study also highlighted Aβ’s role in fueling cognitive deterioration. Some postmortem studies suggested that after 80, Aβ plaques no longer correlate strongly with dementia (Savva et al., 2009; Haroutunian et al., 2008). “In this study you see a real impact of β-amyloid, even in old age,” noted Pieter Jelle Visser, VU University Medical Center in Amsterdam, adding that the finding aligns with another recent study of subjects from the AD Neuroimaging Initiative (ADNI) (Donohue et al., 2017). 

Snitz thinks more insights lie ahead. “We are continuing to follow these oldest-old volunteers and we know biomarker status has changed for many,” she wrote. Visser suggested including neurodegeneration biomarkers beyond hippocampal atrophy, and applying multivariate analyses to continuous measurements, rather than categorizing them as simply positive or negative. This approach could yield more informative, unbiased insights into how cognitive performance and pathology relate to each other, he said. Knopman said that a next important step will be to include tau pathology as a separate marker from hippocampal volume and glucose metabolism, as proposed in the new amyloid/tau/neurodegeneration classification system (Aug 2016 news). “I think it will increase clarity,” he said.—Marina Chicurel

Comments

  1. Zhao and colleagues studied neurocognitive outcomes over ~12 years for 175 persons stratified by baseline PiB-PET (Aβ) status and structural MRI (operationalized by hippocampal volume, HV). The research participants were a subgroup of the Ginkgo Evaluation of Memory (GEM) study, who were enrolled in the subsequent Ginkgo Evaluation of Memory Study (GEMS) Imaging Substudy. This was an elderly and predominantly Caucasian population, with mean age at imaging of 86.0 yrs.

    In this study, the authors classified the persons according to the baseline presence or absence of Ab biomarker positivity by PiB-PET scan (AB+/–), and decreased HV on MRI scans (ND+/–).   They found that persons in each group (AB+/ND+, AB+/ND–, AB–/ND+, and AB–/ND–) had somewhat distinct “typical” cognitive outcomes. Perhaps as to be expected, the “classic AD” AB+/ND+ persons had the worst outcomes in terms of cognitive trajectories, but the AB–/ND+ (suspected non-Alzheimer’s pathology, or SNAP) profile also experienced, with time, a substantial decline in a number of domains, including visual memory.

    There are many strengths to this study, including the outstanding clinical research team, the long follow-up with longitudinal visits featuring an extensive battery of neurocognitive assessments, and the topical relevance of these findings to clinicians and researchers interested in dementia.

    This study adds to a growing appreciation of the importance of non-AD pathophysiology in patients who otherwise may meet the clinical criteria for MCI and subsequently for “Probable AD” according to the McKhann criteria. Notably, a large proportion of the research subjects (20 percent) had the AB–/ND+ biomarker profile that indicates the presence of SNAP, i.e., neurodegenerative disease with no AD. Since the allele frequency of ApoE4 positivity (19 percent) in this sample was around 50 percent higher than the general Caucasian population (almost always higher in research volunteers recruited to study AD!), a study of a more representative epidemiologic sample could find even greater relative impact of SNAP versus pure AD.

    Some of the limitations were as stated trenchantly in the article: "Because the imaging lagged behind the initial cognitive assessment, biomarker status at initial intake was unknown. In addition, ND was operationalized only by HV, simplifying the complexity of ND biomarkers. Finally, the participants were relatively highly educated and mostly of white European descent; the findings may not generalize to other populations."

    In addition, it would be great to know the neuropathologic endpoints (i.e., detailed autopsy results) on as many of the patients as possible. Even lacking those data, we know that many of the AB–/ND+ patients probably had some combinations of hippocampal sclerosis/cerebral age-related TDP-43 pathology with sclerosis (CARTS), primary age-related tauopathy (PART), aging-related tau astrogliopathy (ARTAG), α-synucleinopathy, brain arteriolosclerosis, and/or other cerebrovascular pathologies.  Among the >50 percent of the persons included in this study who had the AD-specific pathophysiologic biomarker (AB+), we can also assume that a great many of these persons also had some combination of these pathologies as well. Whereas they may not constitute “SNAP” according to current definitions, their non-AD brain diseases are definitely relevant to the patients, clinicians, and clinical trials.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Suspected Non-Alzheimer Pathophysiology: It’s Not Exactly a Snap
  2. Staging of Alzheimer’s, the Second: Neurodegeneration Does Not Equal Tauopathy

Paper Citations

  1. . Suspected non-Alzheimer disease pathophysiology--concept and controversy. Nat Rev Neurol. 2016 Feb;12(2):117-24. Epub 2016 Jan 18 PubMed.
  2. . 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. Lancet Neurol. 2016 Sep;15(10):1044-53. Epub 2016 Jul 20 PubMed.
  3. . Heterogeneity in Suspected Non-Alzheimer Disease Pathophysiology Among Clinically Normal Older Individuals. JAMA Neurol. 2016 Oct 1;73(10):1185-1191. PubMed.
  4. . Hypothetical Preclinical Alzheimer Disease Groups and Longitudinal Cognitive Change. JAMA Neurol. 2016 Jun 1;73(6):698-705. PubMed.
  5. . Age-specific population frequencies of cerebral β-amyloidosis and neurodegeneration among people with normal cognitive function aged 50-89 years: a cross-sectional study. Lancet Neurol. 2014 Oct;13(10):997-1005. Epub 2014 Sep 4 PubMed.
  6. . Age, neuropathology, and dementia. N Engl J Med. 2009 May 28;360(22):2302-9. PubMed.
  7. . Role of the neuropathology of Alzheimer disease in dementia in the oldest-old. Arch Neurol. 2008 Sep;65(9):1211-7. PubMed.
  8. . Association Between Elevated Brain Amyloid and Subsequent Cognitive Decline Among Cognitively Normal Persons. JAMA. 2017 Jun 13;317(22):2305-2316. PubMed.

External Citations

  1. GEMS

Further Reading

Papers

  1. . An operational approach to National Institute on Aging-Alzheimer's Association criteria for preclinical Alzheimer disease. Ann Neurol. 2012 Jun;71(6):765-75. PubMed.

Primary Papers

  1. . Amyloid β Deposition and Suspected Non-Alzheimer Pathophysiology and Cognitive Decline Patterns for 12 Years in Oldest Old Participants Without Dementia. JAMA Neurol. 2017 Nov 6; PubMed.