Premium
F5‐01‐03: LONGITUDINAL AMYLOID AND CROSS‐SECTIONAL TAU PET IN DOWN SYNDROME
Author(s) -
Christian Brad T.,
Lao Patrick J.,
Hartley Sigan L.,
Tudorascu Dana L.,
Devenny Darlynne,
Johnson Sterling C.,
Klunk Bill E.,
Handen Benjamin L.
Publication year - 2018
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2018.06.2970
Subject(s) - neuropathology , precuneus , pittsburgh compound b , medicine , biomarker , cohort , amyloid (mycology) , alzheimer's disease , oncology , psychology , pathology , disease , cognition , psychiatry , biology , biochemistry
significant advancements in the search for biomarkers of AD pathology, the search for blood-based biomarkers that can detect and predict AD risk among adults with DS has received less attention. Here we applied our previously established biomarker profile for detecting AD in the general population to a large cohort of adults with DS. Methods: Biobanked plasma samples from adults with DS who have been followed over time were assayed. Proteomics were conducted via electrochemiluminescence from our previously generated algorithm. Support vector machine (SVM) analyses were utilized to create a proteomic risk score taking into account the variable importance of each of the markers specifically for prevalent AD and incident AD. We conducted Cox proportional hazards modeling to examine the relation of the proteomic risk score to onset of dementia in those above the cut point compared with those below the cut point, adjusting for age, sex, level of intellectual disability, race/ethnicity and the presence of an APOE E4 allele. Results: There were 399 participants included in the analysis. Of these 53 were determined to have prevalent dementia, 71 developed dementia and 275 remained nondemented over the 6 year follow-up period. The SVM risk score accurately detected prevalent dementia in DS with an AUC1⁄40.89, sensitivity (SN) 1⁄4 0.83 and specificity (SP) 1⁄4 0.98. The SVM risk score was also highly accurate in predicting incident AD. For incident dementia, the SVM risk score yielded an AUC1⁄40.90, SN1⁄40.80 and SP1⁄40.93. When the SVM-based risk scores for incident AD were entered into Cox proportional models, the HR for incident dementia 1⁄49.89 (95% CI1⁄4 5.816.9). Of note, the predictive algorithm was heavily weighted for inflammatory markers (e.g. eotaxin, TARC, I309, IL10, IL5, TNFa). Conclusions: Overall, the current results provide strong support for the use of our blood-based proteomic profile for both detecting and predicting dementia among adults with DS. As we have proposed for AD in the general population, the goal of the blood-based biomarker is to serve as the first-step in a multi-tiered neurodiagnostic approach.