Premium
Operationalization of the ATN classification scheme in preclinical AD: Findings from EPAD V500.0 data release
Author(s) -
Ingala Silvia,
de Boer Casper,
Masselink Larissa A,
Vergari Ilaria,
Lorenzini Luigi,
Blennow Kaj,
Chetelat Gael,
Perri Carol Di,
Ewers Michael,
Fox Nick C.,
Gispert Juan Domingo,
Molinuevo Jose Luis,
Terrera Graciela Muniz,
Mutsaerts Henri J.M.M.,
Ritchie Craig W.,
Schmidt Mark E.,
Vermunt Lisa,
Waldman Adam,
Wink Alle Meije,
Wolz Robin,
Wottschel Viktor,
Scheltens Philip,
Visser Pieter Jelle,
Barkhof Frederik
Publication year - 2020
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 - 10.1002/alz.037912
Subject(s) - cohort , medicine , psychology
Background The goal of this study is to assess neuropsychological and radiological characteristics according to the ATN scheme in the EPAD Longitudinal Cohort Study (EPAD‐LCS) after establishing data‐driven cut‐offs for classification. Method ATN cut‐off values were determined in a deeply phenotyped cohort of 500 nondemented elderly individuals. Amyloid and p‐tau in the cerebrospinal fluid were quantified with Roche Elecsys assays. Hippocampal volume was calculated using LEAP pipeline. The intersection poin of two‐component Gaussian mixture models was used to establish cut‐off values for CSF amyloid ß42 (A) and p‐tau (T). Age‐adjusted w‐scores were used to establish a cut‐off value for neurodegeneration (N). Only subjects classified as healthy controls (A‐T‐N‐) or in the Alzheimer spectrum (A+T‐N‐, A+T+N, A+T+N+) were included. Multinomial regression was performed to test whether significant differences among the ATN groups could be detected with the EPAD Neuropsychological Examination (ENE) after correcting for age, sex, years of education, and site of data collection. Result Out of the initial sample of 500 study participants, n=38 did not fulfil EPAD inclusion criteria, n=34 did not have CSF or MRI biomarkers, and n=49 were not either healthy controls or in the Alzheimer spectrum and were thus excluded. Cut‐offs were 1025 pg/mL for CSF amyloid ß1,42 (A), 24 pg/mL for CSF p‐tau (T), and age‐adjusted hippocampal volume w‐score below ‐2 standard deviations for neurodegeneration (N). Out of the n=376 sample, 229 individuals were classified as healthy controls (A‐T‐N‐), 102 as showing only amyloid pathology (A+T‐N‐), 45 showing amyloid and tau pathologies (A+T+N‐), and 3 A+T+N+ were eliminated [Table 1]. The ENE battery distinguished healthy controls (A‐T‐N‐) from individuals with preclinical AD (A+T+N‐), and individuals with Alzheimer pathological changes (A+T‐N‐) from individuals with preclinical AD (A+T+N‐). Conclusion We proposed a data‐driven framework of cut‐offs for classifying preclinical AD patients in the ATN scheme. The ENE battery scores were in line with the expected cognitive decline in the AD spectrum and increasing vascular burden was shown with progression along the AD spectrum. Potential applications include the selection of patients for secondary prevention trials, preclinical disease modelling and monitoring responses to disease‐modifying therapies.