Advertisement

PREDICTION OF AMYLOID POSITIVITY WITH DEMOGRAPHIC, APOE AND NEUROPSYCHOLOGICAL INFORMATION IN COGNITIVELY NORMAL OLD ADULTS WITH SUBJECTIVE COGNITIVE DECLINE

      Background

      There are evidences that subjectively experienced cognitive decline even at the stage of normal performance on cognitive tests is associated with an increased risk for future cognitive decline and Alzheimer’s disease (AD) dementia. However, subjective cognitive decline (SCD) is unspecific, and related to various medical, psychiatric, and even cultural conditions. For the purpose of early detection of AD, therefore, identifying the characteristics of SCD at the preclinical stage of AD is very important. We aimed to elucidate the demographic, clinical, neuropsychological, and genetic characteristics associated with beta-amyloid (Aβ) deposition, and to search the best prediction model for Aβ positivity in cognitively normal (CN) adults with SCD.

      Methods

      One-hundred thirty two CN elderly with SCD were recruited from the Korean Brain Aging Study for Early Diagnosis and Prediction of Alzheimer’s Disease (KBASE). The presence of SCD was defined by the answer for a single question: “Do you feel that your cognitive function has declined compared to past?” All the subjects received comprehensive clinical and neuropsychological assessments, apolipoprotein E (APOE) genotyping, 11C-labelled Pittsburgh Compound B (PiB) –PET imaging. Aβ positivity was determined if the SUVR value was > 1.4 in at least one of the four ROIs

      Results

      Univariate analyses showed that Aβ positive group was older and had more APOE e4 carrier than Aβ negative group. In regard of neuropsychological performance, Aβ positive subjects had higher phonemic fluency score than Aβ negative ones. A series of multiple logistic regression analyses showed that the model including age, APOE genotype, and phonemic fluency score together had best predictive ability for Aβ positivity in CN subjects with SCD (prediction accuracy: 87.1%), although the models with each of the three variables or any two of them significantly predict Aβ positivity as well (accuracy: 81.8∼85.6%).

      Conclusions

      Our findings suggest that older age, the presence of APOE e4 allele, and higher phonemic fluency increase the likelihood of preclinical AD in cognitively healthy old adults with subjective cognitive complaint. Higher phonemic fluency may be related to compensatory brain activation for Aβ pathology.
      Table 1Results Obtained From Logistic Regression Analyses Designed to Select Appropriate Models for Amyloid positivity prediction in SCD(+)
      Models-2LLχ2dfpDiagnostic Accuracy (%)Significance test for -2LL Difference
      One-candidate model
       Model C Age103.7215.221<0.00181.8
       Model S Age at onset of SCD107.0711.8810.001
       Model N Phonemic fluency112.236.7210.0183.3
       Model G APOE4+110.468.4910.00383.3
      Two-candidate model
       Model CN Age+ Phonemic fluency92.4426.512<0.00184.1Model CN versus C: p=0.001

      Model CN versus N: p=<0.001
       Model CG Age+ APOE4+95.6423.312<0.00185.6Model CG versus C: p=0.004

      Model CG versus G: p=<0.001
       Model NG Phonemic fluency+ APOE4+101.2217.732<0.00183.3Model NG versus N: p=0.001

      Model NG versus G: p=0.004
      Three-candidate model
       Model CNG Age+ Phonemic fluency+ APOE4+80.9737.983<0.00187.1Model CNG versus CN: p=0.001
      Abbreviations: Apo E4+, apolipoprotein ε4 carrier; SCD, subjective cognitive decline.
      Table 2Finally Selected Logistic Regression Model (model CNG)* for Amyloid positivity Prediction in SCD(+)
      VariablesRegression CoefficientStandard ErrorOdd ratios95% Confidence Intervalp
      Intercept-16.443.590.00<0.001
      Age0.160.041.181.08-1.28<0.001
      Phonemic fluency0.090.021.091.04-1.15<0.001
      ApoE4+2.120.648.322.38-29.040.001
      Abbreviations: Apo E4+, apolipoprotein ε4 carrier; SCD, subjective cognitive decline.