Researchers studying the natural history of Alzheimer’s disease are grappling with a puzzling group of volunteers—amyloid-free and cognitively normal older adults who show other biomarker evidence of neurodegeneration. Initially reported in 2011, sizable numbers of what were, perhaps wrongly, thought to be preclinical AD patients now turn up as having this suspected non-amyloid pathology (SNAP) in two new studies. In a longitudinal analysis of Minnesota seniors, 42 percent of those who eventually racked up high brain amyloid at baseline already had biomarkers suggestive of neurodegeneration even though their brain-amyloid scan was still negative. And a cross-sectional study by researchers in California found that many cognitively normal elderly had neurodegeneration biomarkers in brain regions typically affected by AD, yet lacked Aβ deposition. These California seniors were more likely than those with high brain Aβ to have cerebrovascular disease and perform poorly on cognitive tests. Reported in the October 16 Neurology and October 28 JAMA Neurology, the studies were led by Clifford Jack of Mayo Clinic, Rochester, Minnesota, and William Jagust of University of California, Berkeley, respectively. Together, the findings confirm that SNAP cases are fairly common, and suggest that age-related non-AD conditions can muddle the interpretation of data on AD biomarker progression.

There was no hint of SNAP in a study of Dominantly Inherited Alzheimer Network (DIAN) reported November 5 in the Proceedings of the National Academy of Sciences. Researchers led by John Morris at Washington University School of Medicine, St. Louis, analyzed amyloid and neurodegenerative biomarkers in adults with autosomal-dominant mutations years before they are expected to develop early onset AD. Because the participants are middle-aged, the study steers clear of the comorbidities that confound studies of older cohorts. In this longitudinal study, the researchers consistently found brain amyloid rising first, years before neurodegenerative abnormalities. However, the ordering and extent of neurodegenerative biomarker changes differed by brain region, raising questions as to why some brain structures appear to resist metabolic loss and atrophy better than others.

With the publication several years ago of new AD diagnostic criteria (see Apr 2011 news story), scientists have redoubled their efforts to define the fluid and neuroimaging biomarkers that map the preclinical stage preceding full-blown disease. Jack and colleagues outlined a hypothetical model whereby brain Aβ accumulates early, followed by changes in markers of neuronal metabolism and brain atrophy (see Jan 2010 ARF webinar; Feb 2013 conference story). Recent biomarker studies in healthy elderly seemed broadly consistent with this model, but found, to the surprise of some scientists, that about a quarter of participants had neurodegeneration markers in the absence of brain amyloid (Jack et al., 2012; Knopman et al., 2012).

Because researchers define preclinical AD as a transition from negligible to positive brain amyloid, the current work by Jack and colleagues addresses what happens in the brain at the onset of AD pathology. They measured amyloid deposition in healthy elderly people in the Mayo Clinic Study of Aging using positron emission tomography (PET) with the amyloid tracer Pittsburgh Compound B (PIB). They found that 123 of 207 participants were amyloid-free at baseline; of those, 26 became amyloid-positive during an average 1.3-year follow-up. Curiously, 11 of the 26 (42 percent) already had an established marker of neurodegeneration—low glucose metabolism or small hippocampi—at baseline, when brain amyloid still fell within the normal range.

In the JAMA Neurology study, first author Miranka Wirth and colleagues analyzed 72 nondemented older adults in the Berkeley Aging Cohort. Twenty-nine people (40 percent) had at least one abnormal neurodegenerative biomarker (hippocampal volume, glucose metabolism, or gray-matter thickness), and among those, 19 people were amyloid-negative by PIB PET. Those 19 amyloid-free neurodegeneration cases comprised 26 percent of the cohort—a figure similar to those in the 2011 studies that first reported SNAP.

Several issues make the SNAP group difficult to analyze. First, PET tracers only measure fibrillar Aβ. “You have to wonder if people have forms of amyloid we cannot measure,” Jagust said. “Soluble oligomeric Aβ could be causing the neurodegeneration we are calling ‘amyloid-negative.’” The absence of SNAP in DIAN would argue against that idea, since Aβ drives pathology in mutation carriers and yet none show the “neurodegeneration-first” phenotype. Another potential complication stems from the way cutoffs are defined, and whether amyloid burden should be measured as trajectories instead of binary positive/negative categories. While some cognitively normal elderly accumulate enough Aβ to cross a predetermined cutoff that defines amyloid positivity, others may have lower yet fast-rising amyloid levels. “Just because they didn’t hit some threshold doesn’t mean the amyloid isn’t causing trouble,” Jagust said. “We don’t know that the point at which amyloid causes problems is the point we define as ‘positive.’”

Second, the SNAP category may include people who have other non-AD pathologies before starting to accumulate brain amyloid. “The older the cohort, the less likely you are to have pure AD, and the more likely you are to have AD arising in conjunction with, or maybe after, other degenerative pathologies,” Jack told Alzforum, noting that the neurodegeneration biomarkers are not specific for AD. In support of this, a recent paper by Jagust and colleagues showed that gray-matter atrophy in cognitively normal seniors is much more closely associated with age than with amyloid (Oh et al., 2013). “There are likely many things driving what we call ‘neurodegeneration markers,’” Jagust said. “One is probably amyloid, another may be vascular disease, or other molecular processes associated with aging.”

Both studies found a number of SNAP cases even though they used different methods for collecting and analyzing data, Jagust said. His group used the distribution volume ratio (DVR) PIB PET approach, which requires longer scans but is believed to be quantitatively more robust than the quicker standardized uptake value ratio (SUVR) method used by Jack and colleagues. Jack’s team analyzed more participants, which helps clean up the noisy SUVR data. In addition, the two groups use different ways of defining amyloid positivity. Jack’s group defines the threshold based on AD cases, whereas Jagust and colleagues establish cutpoints as standard deviations above amyloid levels in healthy young adults. Despite these procedural differences, “the results are pretty consistent,” Jagust said. “SNAP seems to be a real phenomenon, at least across two cohorts.”

What defines this unusual group and how fast its members decline remains unclear. While some analyses suggest that people with SNAP deteriorate no faster than normal seniors who lack amyloid or neurodegenerative markers (see Vos et al., 2013), others paint a more confusing picture (see Aug 2013 conference story).

SNAP does not show up in familial AD. As reported in PNAS this week, first author Tammie Benzinger and colleagues analyzed 137 autosomal-dominant mutation carriers and 92 noncarriers from DIAN. Broadly speaking, their study confirmed the ordering of biomarkers proposed for sporadic AD—that is, amyloid deposition showing up early, followed by metabolic loss, and later, cortical thinning. “Everywhere we looked, we detected amyloid first,” Benzinger said. However, certain brain areas did not follow expected patterns of hypometabolism and atrophy. For instance, the hippocampus loses massive numbers of cells, despite having low amyloid burden compared to the Aβ amounts measured elsewhere. Other areas accumulate whopping amounts of amyloid yet do not succumb to hypometabolism until much later. Most importantly, the study confirms that “for cortical structures, amyloid deposition really does occur 15 to 20 years before symptoms,” she said. “That gives us a good therapeutic window.”—Esther Landhuis


  1. The significance of these findings, which now validate the phenomenon of SNAP, cannot be overstated. In 2009, based on the data then available, I suggested that amyloid is not the cause of AD, but a cause of AD (Pimplikar, 2009), and that amyloid-independent mechanisms must also be at work. Later, together with Drs. Nixon, Robakis, Shen, and Tsai, I presented multiple, amyloid-independent mechanisms of neurodegeneration that could be operative in AD (Pimplikar et al., 2010) and, therefore, potential drug targets. The current observations that older, cognitively normal individuals and patients diagnosed with MCI/AD exhibit neurodegeneration in the absence of amyloid plaques (see Aug 2013 news) can be seen as a clinical validation of the “amyloid-independent” pathogenesis of AD (Petersen et al., 2013). There is now enough clinical evidence to conclude that amyloid is neither sufficient (based on cognitively normal PiB-positive individuals) nor necessary (MCI/AD patients with SNAP) to cause AD (there is no clinical evidence that oligomeric Aβ, invisible to PiB imaging, is responsible for AD and the preclinical studies on Aβ oligomers are wrought with artifacts, see Oct 2011 Webinar).

    Such a conclusion, uncomfortable as it might seem, has important implications for AD drug trials. Despite the multiple failures of amyloid-focused drugs in MCI/AD patients, leaders in the field have doubled down on the amyloid-centric approach and a significant amount of the financial resources of the NIH have been directed toward long-term trials in asymptomatic individuals (see Sep 2013 news story). On one hand, the knowledge that neurodegeneration in AD patients (or future AD patients) with SNAP could be due to amyloid-independent mechanisms should increase the success of amyloid-focused drugs by excluding such patients. On the other hand, since the SNAP cohort of AD patients is estimated to be between 17 and 29 percent (Petersen et al., 2013), between 1 million and 1.5 million Americans with AD will need alternative therapeutic intervention.

    If nothing else, these two papers once again underscore the urgent need to evaluate AD pathogenesis through an amyloid-independent framework and the fact that resources are required to rigorously test alternative pathogenic mechanisms and therapies based on such hypotheses. We must look beyond amyloid and diversify our AD drug target repertoire.


    . Reassessing the amyloid cascade hypothesis of Alzheimer's disease. Int J Biochem Cell Biol. 2009 Jun;41(6):1261-8. PubMed.

    . Amyloid-independent mechanisms in Alzheimer's disease pathogenesis. J Neurosci. 2010 Nov 10;30(45):14946-54. PubMed.

    . Criteria for mild cognitive impairment due to alzheimer's disease in the community. Ann Neurol. 2013 May 20; PubMed.

  2. I suggest that the E693delta (Osaka) mutation in APP may be pertinent to the SNAP cases (e.g., Tomiyama et al., 2008; Shimada et al., 2011).


    . A new amyloid beta variant favoring oligomerization in Alzheimer's-type dementia. Ann Neurol. 2008 Mar;63(3):377-87. PubMed.

    . Clinical course of patients with familial early-onset Alzheimer's disease potentially lacking senile plaques bearing the E693Δ mutation in amyloid precursor protein. Dement Geriatr Cogn Disord. 2011;32(1):45-54. PubMed.

Make a Comment

To make a comment you must login or register.


Webinar Citations

  1. Together at Last, Top Five Biomarkers Model Stages of AD

News Citations

  1. HAI—Sharper Curves: Revamping a Biomarker Staging Model
  2. Suspected Non-Amyloid Pathology (SNAP)—Not an Open or Shut Case

Paper Citations

  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.
  2. . Short-term clinical outcomes for stages of NIA-AA preclinical Alzheimer disease. Neurology. 2012 May 15;78(20):1576-82. PubMed.
  3. . Association of Gray Matter Atrophy with Age, β-Amyloid, and Cognition in Aging. Cereb Cortex. 2013 Feb 6; PubMed.
  4. . Preclinical Alzheimer's disease and its outcome: a longitudinal cohort study. Lancet Neurol. 2013 Oct;12(10):957-65. PubMed.

Other Citations

  1. Apr 2011 news story

External Citations

  1. DIAN

Further Reading


  1. . Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010 Jan;9(1):119-28. PubMed.
  2. . Tracking pathophysiological processes in Alzheimer's disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol. 2013 Feb;12(2):207-16. PubMed.
  3. . Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 2011 May;7(3):280-92. Epub 2011 Apr 21 PubMed.
  4. . Alzheimer's disease neurodegenerative biomarkers are associated with decreased cognitive function but not β-amyloid in cognitively normal older individuals. J Neurosci. 2013 Mar 27;33(13):5553-63. PubMed.

Primary Papers

  1. . Regional variability of imaging biomarkers in autosomal dominant Alzheimer's disease. Proc Natl Acad Sci U S A. 2013 Nov 19;110(47):E4502-9. Epub 2013 Nov 5 PubMed.
  2. . Amyloid-first and neurodegeneration-first profiles characterize incident amyloid PET positivity. Neurology. 2013 Nov 12;81(20):1732-40. PubMed.
  3. . Amyloid and neurodegeneration: Converging and diverging paths. Neurology. 2013 Nov 12;81(20):1728-9. PubMed.
  4. . Associations between Alzheimer disease biomarkers, neurodegeneration, and cognition in cognitively normal older people. JAMA Neurol. 2013 Dec;70(12):1512-9. PubMed.