P3‐081: Application of Clinical‐Based Indices on Identification of Genetic Markers for Late‐Onset Alzheimer’s Disease

P-values for selection of single nucleotide polymorphisms (SNPs) have suffered from high false positive and false negative results. This study was aimed to compare the application of clinical-based index and p-value for marker selection. This case-control study included 713 late-onset Alzheimer’s disease (LOAD) cases and controls aged 65 or order (2007 to 2010) from three hospitals in northern Taiwan. Clinical-based indices (including sensitivity, specificity, positive predictive value, negative predictive value, and accuracy) were used to select SNPs at stage 1, which then genotyped in another dataset (stage 2). A point-based genetic model including the SNP selected by clinical-based indices, apolipoprotein E (APOE) e4 status, and important confounders was established for clinical implications. Four SNPs on CPXM2, APOC1, and ZNF521 gene were identified for LOAD by both traditional P-value (if no correction for multiple tests) and clinical-based indices. After correction for multiple tests, no SNPs were identified by traditional P-value. Simultaneous testing of APOE e4 and the SNP on APOC1 (SNP with the best performance on clinical-based indices) significantly improved the low sensitivity of APOE e4 (0.78). A point-based genetic model including these 2 SNPs and important confounders were constructed. Compared with elders with low risk score (0-3), elders belong to moderate (score=4-6) and high risk (score=7-10) groups had significantly increased risk of LOAD (adjusted odds ratio=8 and 52, respectively; Ptrend < 0.0001). Clinical-based indices allow us to identify markers with moderate effect and are useful for translation of genetic testing for clinical and public health implications.