Mielke M, Wiste H, Wiegand S, Knopman D, Lowe V, Roberts R, Swenson-Dravis D, Boeve B, Petersen R, Jack C.
Predictors of Amyloid Accumulation in a Population-Based Study of Cognitively Normal Elderly Controls.
Human Amyloid Imaging Abstract. 2012 Jan 1;
Background: Secondary Alzheimer’s disease (AD) prevention trials in preclinical subjects are now being designed. Documentation of brain amyloidosis for subject enrollment is a requirement of some trials. Approximately0% of elderly individuals are expected to be amyloid negative. The identification of inexpensive and non-invasive screening variables that could help predict which individuals have significant amyloid accumulation would greatly reduce screening costs for preclinical AD trials.
Methods: A population-based sub-sample of83 cognitively normal (CN) individuals, aged0-93, from the Mayo Clinic Study of Aging underwent PIB-PET imaging. Logistic regression was used to determine whether age, APOE genotype, family history, or cognitive performance increased the odds of having PIB SUVR>1.5. Area under the receiver operating characteristic curve (AUROC) was used to evaluate discrimination between PIB positive and negative subjects. Positive (PPV) and negative (NPV) predictive value was defined based on an estimated probability >0.50 who were PIB>1.5.
Results: Of the83 CN individuals1 (31%) had PIB>1.5. Each five-year increase in age (OR.6% CI:.3,.9) and presence of an APOE E4 allele (OR.7% CI:.3,.7) were independently associated with PIB>1.5. Adding family history of dementia/AD, subjective memory complaints, and cognitive performance did not appreciably improve the classification of persons with PIB>1.5 (AUROC=0.69, PPV=60%, and NPV=73% with age and APOE genotype vs. AUROC=0.71, PPV=61%, and NPV=74% with additional variables).
Conclusion: Age and APOE genotype are useful predictors of PIB>1.5. However, cognitive performance in any domain and subjective memory complaints did not improve classification. Prediction using age and APOE genotype was only fair. However, a PPV of0% versus1% (expected baseline rate) means these two inexpensive and non-invasive measures could reduce, by half, the number of CN subjects that must be screened in order to identify a given number of amyloid positive subjects.