23 September 2013. Proposed preclinical stages for Alzheimer’s disease work well to predict who is most likely to progress to AD, according to a paper released online by Lancet Neurology on September 4. Researchers led by Stephanie Vos and Pieter Jelle Visser at Maastricht University in the Netherlands and Anne Fagan and others at Washington University in St. Louis, Missouri, analyzed data from 311 cognitively normal elderly people seen at St. Louis’s Knight Alzheimer’s Disease Research Center. Those with the greatest biomarker evidence of Alzheimer’s pathology were most likely to develop dementia over the next four years, the authors report. Vos previously presented these data at the Alzheimer's Association International Conference (AAIC), held July 14-18 in Boston, Massachusetts (see ARF related news story)
In the current model, which was commissioned by the National Institute on Aging and the Alzheimer’s Association (see ARF related news story), cognitively normal people without biomarker evidence of AD are “stage 0,” while those with evidence of amyloid are stage 1, and those with evidence of both amyloid and a neurodegeneration marker are stage 2. People at stage 3 show subtle cognitive deficits in addition to biomarkers, although their Clinical Dementia Rating (CDR) remains normal at 0. The Vos et al. study used CSF biomarkers.
By the end of the study, 56 percent of participants who were initially classified as stage 3 had progressed to symptomatic AD. This was defined as a CDR of at least 0.5 plus signs of dementia believed to be due to AD as determined by a clinician, in agreement with published criteria (see McKhann et al., 1984). These signs might include abnormalities in mood, language, and behavior. In contrast, only 26 percent of people at stage 2, 11 percent at stage 1, and 2 percent of those at stage 0 developed dementia during follow-up. When plotted against incidence of dementia due to Alzheimer’s, the progression curves of people at stage 0, 1, 2, 3, or a different group called SNAP (suspected non-amyloid pathology defined by amyloid negativity in the presence of biomarker positivity for neurodegeneration), separated completely over the course of up to 14 years of follow-up. There was no overlap between any of the curves.
The paper’s terminology sparked controversy that reflects the field’s struggle over what exactly to call people at preclinical, or very early and subtle cognitive, stages of Alzheimer’s, and how to separate them from people with very subtle signs of diseases other than AD. In a comment released along with the paper, Ron Petersen at the Mayo Clinic in Rochester, Minnesota, questioned the authors’ use of the term “symptomatic AD,” noting that the stage following preclinical disease is usually designated as “mild cognitive impairment (MCI) due to AD” or “prodromal AD.” Most other biomarker studies “use some form of MCI as the intermediate clinical category,” Petersen wrote.
In the authors’ reply, published open-access on September 6, they agree that an MCI stage precedes AD. Some participants indeed developed cognitive impairment and received a CDR rating of 0.5, but did not show signs of dementia and thus were not counted as having progressed to AD, Fagan told Alzforum. Because the CDR is a staging rather than diagnostic tool, a CDR of 0.5 encompasses some people with very mild dementia as well as some with MCI due to other underlying conditions, she noted. The authors chose to use dementia, rather than MCI, as the endpoint. “We are interested in knowing whether preclinical stages predict AD dementia, not how well they predict cognitive impairment of unknown etiology,” Fagan explained. “Had we used progression to MCI as the outcome measure, the proportion of people in the different stages who progressed would likely have been higher.”
Regardless of this debate over the precise delineation between MCI and AD, all the researchers agree that Alzheimer’s disease begins long before clinical symptoms and that the proposed research criteria are being largely borne out as they are tested in prospective cohorts. This study and others (see, e.g., Knopman et al., 2012) “provide strong support for the validity of the construct of preclinical AD,” Petersen wrote in his comment. The article and Petersen’s comment appear in the October print issue of the journal, while the authors’ reply will run in the December issue. —Madolyn Bowman Rogers.
Vos SJ, Xiong C, Visser PJ, Jasielec MS, Hassenstab J, Grant EA, Cairns NJ, Morris JC, Holtzman DM, Fagan AM. Preclinical Alzheimer’s disease and its outcome: a longitudinal cohort study. Lancet Neurol. 2013 Sep 3. Abstract
Petersen RC. Do preclinical Alzheimer’s disease criteria work? Lancet Neurol. 2013 Sep 3. Abstract
Fagan AM, Vos SJ. Preclinical Alzheimer’s disease criteria. Lancet Neurol. 2013 Sep 6. Abstract