When judging the utility of cerebrospinal fluid biomarkers in Alzheimer disease, it may be less a question of if, but rather when and in whom. In this week’s Journal of Neuroscience, scientists report analysis of CSF biomarkers and brain atrophy measures and find that the two, to a large degree, do not track with each other in AD and mild cognitive impairment (MCI). However, the same group reported online January 4 in the journal Cerebral Cortex that CSF Aβ levels do associate with one-year brain atrophy in cognitively normal seniors. And in this month’s Archives of Neurology, another study of healthy elders finds that CSF biomarker change correlates with cognitive decline. Together, the studies seem to suggest that CSF markers reflect ongoing pathology early in the disease, but less so in later stages.

At this point, the field acknowledges low CSF Aβ, elevated CSF tau and phospho-tau, and brain atrophy as reputable readouts of AD pathology. Along the pathogenic cascade, CSF biomarker change is typically seen as an early event and atrophy a late event. The Journal of Neuroscience study, led by Anders Dale of the University of California, San Diego, challenges this assumption, at the very least suggesting it is too simplistic. First author Anders Fjell, Ullevaal University Hospital, Oslo, Norway, and colleagues explored the relationship between CSF and brain atrophy, measured by magnetic resonance imaging (MRI), in 307 volunteers over one or two years. The subjects (105 controls, 175 MCI, 90 AD) were participants in the Alzheimer’s Disease Neuroimaging Initiative, a large-scale effort to validate and standardize the various fluid and imaging biomarkers in AD (see ARF related news series).

Several key findings emerged from the reams of CSF measurements and MRI brain images in this data-rich paper. “We saw that even MCI patients with normal CSF biomarker levels had much more brain atrophy than controls,” Fjell said, noting that controlling for CSF biomarkers also could not account for differences in baseline MRI measures between the patient groups. These data show that “CSF and MRI biomarkers are, to a substantial degree, independent of each other.” As such, the two measures are complementary and can inform one another in AD diagnosis and prognosis, Fjell said. He also noted a potentially discouraging ramification for AD drug trials. “These data suggest that even if you’re able to change the CSF Aβ and tau levels back to normal, you may not affect the degree of atrophy,” Fjell said. “That’s the gist of the baseline data.”

When assessed longitudinally, CSF and atrophy measures had a more noticeable correlation. These associations appeared in a wide range of brain areas, not just those implicated in AD. Previous studies had shown a relationship between longitudinal CSF and structural MRI (e.g., Hampel et al., 2005; de Leon et al., 2006), but only looked at hippocampal volume, for instance. The current study examined 30 regions of interest throughout the brain. The researchers also found that baseline MRI measures correlated with clinical change better than did CSF biomarkers. Overall, “the results indicate that morphometric changes in MCI and AD are not secondary to CSF biomarker changes,” the authors write. The apparent independence of MRI and CSF biomarkers suggests to Fjell that “there must be some events that cause atrophy but are not related to CSF biomarkers.” He speculates that people with CSF biomarker changes and normal MRI measures, and those with normal CSF measures and greater atrophy, may, in fact, represent different patient groups. “We may be able to distinguish subtypes of neurodegenerative disease in a better way than we can do today,” he said, noting that additional longitudinal follow-up would be needed to address this.

In an e-mail to ARF, Agneta Nordberg of the Karolinska Institute in Stockholm, Sweden, wrote that further study is required to show how CSF biomarkers change with disease progression. She and colleagues have five-year brain imaging data from positron emission tomography (PET) scans showing stable amyloid levels in MCI and AD patients. “The MCI patients with no CSF Aβ42 changes but with atrophy would have been interesting to study with amyloid imaging as well,” she wrote, noting that these patients might represent prodromal non-AD dementia cases. (See full comment below; for review on use of PET in AD, see Nordberg et al., 2010.)

At any rate, timing appears key in assessing the value of CSF biomarkers. In a study of 71 cognitive normal seniors published last month in Cerebral Cortex, Dale and colleagues report that CSF Aβ42 levels did associate closely with one-year atrophy in a variety of brain areas, including, but not limited to, those affected in early AD (Fjell et al., 2010).

The Archives of Neurology study also examined healthy elderly, specifically whether changes in CSF biomarker levels over time correlate with worsening cognitive performance. Previous reports suggest that CSF biomarkers can predict conversion from MCI to AD (Hansson et al., 2006; Mattsson et al., 2009 and ARF related news story), but few have examined the effectiveness of these markers in presymptomatic individuals. Led by senior investigator Elisabet Londos and first author Erik Stomrud of Lund University in Malmo, Sweden, the current study focused on this so-called pre-MCI stage. The researchers analyzed 37 cognitively healthy seniors, doing CSF measurements (total tau, hyperphosphorylated tau, Aβ1-42) and cognitive assessments at baseline and at follow-up four years later. Participants whose CSF Aβ1-42 levels dropped at least 15 percent over the study’s duration did worse on the delayed word recall portion of the ADAS-cog compared to people with no or milder CSF Aβ changes. Those with a 20 percent or greater increase in CSF levels of hyperphosphorylated tau protein 181 had poorer performance on a test of cognitive speed. “Changes in CSF biomarker levels may identify individuals experiencing pre-MCI and possible early neurodegenerative processes of AD,” the authors write.

Taken together, the new studies appear consistent with a “Personal View” essay in last month’s Lancet Neurology by Clifford Jack, Mayo Clinic, Rochester, Minnesota, which synthesizes the burgeoning literature on AD biomarkers and proposes a model for how the various fluid and imaging measures fit into the pathogenic process (Jack et al., 2010 and ARF Live Discussion).—Esther Landhuis


  1. This is a study of 105 normal controls, 175 MCI, and 90 AD patients followed for on to two years. CSF biomarker and brain atrophy were used as predictive parameters for clinical changes in AD. Surprisingly the authors reported that MCI patients with CSF levels of Aβ42 comparable controls and of CSF tau below controls showed more atrophy than controls. The authors concluded that morphometry predicts cognitive progression better than CSF biomarkers.
    I think it would have been interesting to know the mean ages of these patients (only age range is given).

    CSF biomarkers, especially Aβ42, are considered to be (together with brain amyloid imaging by PET) comparable to atrophy changes. However, when studying the progression of the AD, disease atrophy as well as cerebral glucose metabolism better correlate with cognitive decline compared to CSF biomarkers. CSF Aβ42 could be considered as a biomarker to detect prodromal AD (preclinical AD), but there is no data so far that convincingly show changes in CSF Aβ levels with progression of the disease. Long-term follow-up of PIB amyloid imaging (we have data up to five years) show stable amyloid levels in AD and MCI. We still have to perform more studies to obtain a deeper understanding of how these different biomarkers, for example CSF Aβ42, could be used as early biomarkers.

    MCI patients with no changes in CSF Aβ42, but with atrophy, would have been interesting to study also with PIB imaging, and one can also question whether these MCI patients with atrophy but normal CSF Aβ42 could represent prodromal non-AD. I have examined several AD cases who have normal Aβ42 in CSF but high uptake of PIB in brain. These patients do very often show very little atrophy in the earlier stage of the disease.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Multi-Paper Alert: More Data That Brain Amyloid Is Bad for You

Webinar Citations

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

Paper Citations

  1. . Correlation of cerebrospinal fluid levels of tau protein phosphorylated at threonine 231 with rates of hippocampal atrophy in Alzheimer disease. Arch Neurol. 2005 May;62(5):770-3. PubMed.
  2. . Longitudinal CSF and MRI biomarkers improve the diagnosis of mild cognitive impairment. Neurobiol Aging. 2006 Mar;27(3):394-401. PubMed.
  3. . The use of PET in Alzheimer disease. Nat Rev Neurol. 2010 Feb;6(2):78-87. PubMed.
  4. . Brain atrophy in healthy aging is related to CSF levels of Aβ1-42. Cereb Cortex. 2010 Sep;20(9):2069-79. PubMed.
  5. . Association between CSF biomarkers and incipient Alzheimer's disease in patients with mild cognitive impairment: a follow-up study. Lancet Neurol. 2006 Mar;5(3):228-34. PubMed.
  6. . CSF biomarkers and incipient Alzheimer disease in patients with mild cognitive impairment. JAMA. 2009 Jul 22;302(4):385-93. PubMed.
  7. . Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010 Jan;9(1):119-28. PubMed.

Other Citations

  1. ARF related news series

External Citations

  1. Alzheimer’s Disease Neuroimaging Initiative

Further Reading


  1. . Brain atrophy in healthy aging is related to CSF levels of Aβ1-42. Cereb Cortex. 2010 Sep;20(9):2069-79. PubMed.
  2. . Association between CSF biomarkers and incipient Alzheimer's disease in patients with mild cognitive impairment: a follow-up study. Lancet Neurol. 2006 Mar;5(3):228-34. PubMed.
  3. . Hypothetical model of dynamic biomarkers of the Alzheimer's pathological cascade. Lancet Neurol. 2010 Jan;9(1):119-28. PubMed.

Primary Papers

  1. . Correlation of longitudinal cerebrospinal fluid biomarkers with cognitive decline in healthy older adults. Arch Neurol. 2010 Feb;67(2):217-23. PubMed.
  2. . CSF biomarkers in prediction of cerebral and clinical change in mild cognitive impairment and Alzheimer's disease. J Neurosci. 2010 Feb 10;30(6):2088-101. PubMed.