It is generally estimated that 10–20% of participants in AD trials do not have AD. Without a validated antemortem diagnostic test, and particularly when site expertise may vary substantially, identification of subjects according to standard clinical and psychometric criteria is imperfect, leading to dilution of observable treatment effects on the disease. While not yet done for pivotal trials, addition of a biomarker assessment would be expected to significantly reduce the diagnostic inaccuracy at enrollment. For example, requiring an amyloid signal by amyloid PET scanning or low CSF Aβ42
would reduce the number of individuals with cognitive symptoms caused by conditions other than AD enrolling in trials (Aisen, et al., 2010
As drug development programs move into the pre-dementia population, this issue becomes much more important. Mild cognitive impairment is a heterogeneous clinical syndrome, with 30–40% of individuals amyloid-negative and not destined for Alzheimer dementia. Predictors of progression include APOE genotype, cognitive and clinical scores, imaging measures and cerebrospinal fluid markers; these can be used to enrich a population of individuals with MCI for rapid progression (Aisen, et al., 2010
,Petersen, et al., 2010
). Particularly for anti-amyloid therapeutic programs, it seems appropriate to select individuals for trials on the basis of amyloid PET imaging and/or low CSF Aβ42
; presumably, amyloid biomarkers not only enrich for progression, but also for potential response to anti-amyloid intervention.
APOE genotyping can also be used for selecting pre-dementia subjects more likely to progress to Alzheimer dementia. An advantage of this method is its low cost. However, selecting subjects on the basis of genotype excludes the large portion (30–50%) of individuals with AD who do not carry the ε4 allele. Further, it may raise regulatory difficulties in late stage development, as it would be necessary to establish efficacy in ε4 carriers as well as lack of efficacy in non-carriers.