16 Dec 2011

A popular model of Alzheimer’s disease progression proposes an ordered sequence by which the main biomarkers change: β amyloid first, then tau, followed by hippocampal volume and cognitive function (see ARF Webinar). Researchers led by Cliff Jack at the Mayo Clinic in Rochester, Minnesota, have now provided some evidence in support of this hypothetical cascade. The data, first presented at the 2011 Alzheimer’s Association International Conference in Paris, France, and published online August 8, are now in print in the December Archives of Neurology.

To validate the paradigm, Jack and colleagues analyzed data from about 400 participants in the Alzheimer’s Disease Neuroimaging Initiative whose average age was in the late seventies. Roughly half the participants had mild cognitive impairment (MCI), a quarter had AD, and the rest were cognitively normal. The researchers focused on three biomarkers: Aβ42 levels in cerebrospinal fluid (CSF); CSF total tau levels; and hippocampal volume (adjusted for intracranial volume) as measured by structural MRI. For each biomarker, the authors used a cutoff value to score the trait as either normal or abnormal in each participant. They took cutoff values for Aβ42 (192 pg/mL) and total tau (92 pg/mL) from published work (see ARF related news story on Shaw et al., 2009 and ARF related news story on De Meyer et al., 2010). For adjusted hippocampal volume, the authors established a cutoff value based on a cohort of about 50 Mayo Clinic patients whose AD status was verified by autopsy.

As expected, in the AD group a greater proportion of participants had abnormal scores for each biomarker, and the average score drifted further from the norm than in the MCI group. Ditto for people with MCI compared to controls, confirming that all three biomarkers progress along with the disease. However, in all three groups, many participants had abnormal Aβ scores, while fewer people approached cutoff values for total tau and hippocampal volume, supporting the hypothesis that Aβ changes first of those three.

About three-quarters of the participants provided samples at a one-year follow-up appointment. The authors report that the proportion of people with abnormal Aβ42 scores did not change in any group, nor did the average Aβ42 level. This jibes with the model’s prediction that β amyloid changes peak early in the disease course, well in advance of any cognitive symptoms. By contrast, the proportion of people with abnormal tau scores, as well as their average tau level, jumped in the cognitively normal group alone, showing that this biomarker is dynamic only in presymptomatic people. The proportion of people with abnormal hippocampal volumes grew only in the MCI group. Tau moves before atrophy, but hippocampal volume catches up during the MCI phase, the authors conclude. Although this study supports the proposed biomarker progression model, the authors note that final validation awaits long-term longitudinal studies. Other studies testing the staging model are ongoing and may sharpen it in time.—Madolyn Bowman Rogers