A combination of cerebrospinal fluid biomarkers and simple cognitive testing identified stages of preclinical Alzheimer’s that were associated with cognitive decline and death over a decade of follow-up in a prospective, longitudinal study.
Preclinical disease was present in 31% of adults aged 65 years or older who were living independently in the community and was a reliable predictor of progression. The findings suggest that preclinical staging is not only possible, but could be a useful adjunct for stratifying research populations in therapeutic trials, according Dr. Stephanie J. Vos, the lead investigator (Lancet Neurol. 2013 Sept. 4 [doi:10.1016/S1474-4422(13)70194-7]).
The study aimed to identify the prevalence and long-term outcomes of preclinical Alzheimer’s disease in elderly subjects who were cognitively normal at baseline. Dr. Vos, of the Alzheimer’s Center Limburg at Maastricht (the Netherlands) University, and her colleagues used the combination of biomarkers and cognitive testing to define preclinical stages similar to those recently proposed by the Preclinical Working Group of the National Institute on Aging and the Alzheimer’s Association. These criteria propose three progressive stages for cognitively normal subjects:
• Stage 1, cognitively normal individuals with abnormal amyloid markers.
• Stage 2, abnormal amyloid and neuronal injury markers.
• Stage 3, abnormal amyloid and neuronal injury markers with subtle cognitive changes.
Dr. Vos’ study involved 311 subjects who were cognitively normal at baseline. They underwent lumbar puncture to ascertain cerebrospinal fluid levels of beta-amyloid-42 (Abeta-42), total tau, and phosphorylated tau. They also completed cognitive testing with the Clinical Dementia Rating (CDR) scale and Mini Mental State Exam (MMSE). Each year thereafter, subjects had additional cognitive testing. The primary outcome was the proportion of patients in each preclinical stage at baseline. Secondary outcomes were progression of cognitive decline and mortality.
CSF samples were dichotomized as normal or abnormal based on a level that the investigators determined. The cutoff values for abnormal biomarker measurements that best differentiated subjects with no baseline memory deficits from those in a separate cohort with symptomatic Alzheimer’s were Abeta-42 levels less than 459 pg/mL, total tau levels greater than 339 pg/mL, and phosphorylated tau levels greater than 67 pg/mL.
A CDR sum of boxes score of 0 was considered normal memory; scores of 0.5 or higher during follow-up indicated progression to symptomatic Alzheimer’s.
Subjects were stratified according to a combination of memory scores and biomarkers. Normal subjects had no memory impairment and normal biomarkers. Subjects were classified as stage 1 if only their Abeta-42 was abnormal. Stage 2 patients had abnormal Abeta-42and abnormal total or phosphorylated tau levels. Stage 3 subjects had abnormal biomarkers plus memory impairment equal to 0.5 on the CDR. Those in stage 1, 2, or 3 were considered to have preclinical Alzheimer’s.
Those who had normal Abeta-42 but abnormal tau – a marker of neuronal injury – were considered to have suspected non-Alzheimer’s pathophysiology (SNAP), regardless of their baseline memory score.
Patients were a mean of 73 years old at baseline. They had a mean MMSE score was 28.9 and a mean CDR sum of boxes score of 0.03. One-third (34%) were positive for the high-risk apolipoprotein (APOE) epsilon-4 allele.
At baseline, 129 (41%) were classified as normal; 47 (15%) as stage 1; 36 (12%) as stage 2; 13 (4%) as stage 3; and 72 (23%) as being in the SNAP group. The remaining 14 (5%) were unclassified.
Preclinical Alzheimer’s (stages 1-3) was significantly more prevalent among those older than 72 years than in those younger (37% vs. 26%), and in APOE epsilon-4 carriers than in noncarriers (47% vs. 23%).
At 5 years, 110 subjects were available for follow-up; 14 were available at 10 years. By the end of the study, 20 subjects had died.
After a median follow-up of 4 years, progression had occurred in 2% of the normal group; 13% of those in stage 1; 25% of those in stage 2; 54% of those in stage 3; 6% of those in the SNAP group, and 29% of those in the unclassified group, for a total of 32 subjects.
Of those who progressed, symptomatic Alzheimer’s with a CDR sum of boxes score of 0.5 occurred in 22 (69%), CDR 1 symptomatic disease developed in 6 (19%), and CDR 2 symptomatic disease arose in 4 (13%).
Interestingly, the authors noted, "neither age (younger than 72 years vs. 72 years or older) nor APOE genotype predicted the rate of decline," although these subanalyses had limited statistical power because of the small sample sizes. "Although APOE epsilon-4 is often a good predictor of cognitive decline in unselected populations, the absence of its prognostic utility in individuals with AD pathological abnormalities is consistent with findings from previous studies."