Premium
Ultrasensitive blood biomarkers to predict cognitive decline and diagnose Alzheimer’s disease in the absence of AT(N) classification as the reference standard
Author(s) -
Simrén Joel,
Leuzy Antoine,
Karikari Thomas K,
Hye Abdul,
Mattsson Niklas,
Hansson Oskar,
Schöll Michael,
Mecocci Patrizia,
Vellas Bruno,
Tsolaki Magda,
Kloszewska Iwona,
Soininen Hilkka,
Lovestone Simon,
Aarsland Dag,
Westman Eric,
Blennow Kaj,
Zetterberg Henrik,
Ashton Nicholas J
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.041808
Subject(s) - biomarker , cognitive decline , dementia , clinical dementia rating , medicine , oncology , cohort , positron emission tomography , alzheimer's disease neuroimaging initiative , pittsburgh compound b , alzheimer's disease , neurodegeneration , pathology , disease , nuclear medicine , biology , biochemistry
Background In recent years, blood biomarkers for Alzheimer’s disease (AD) have attracted much interest. Technological advances have allowed quantification of markers for neurodegeneration (NfL), neurofibrillary tangles (p‐tau181) and amyloid‐β pathology (Aβ42, Aβ42/Aβ40), which perform well in cohorts stratified by cerebrospinal fluid and positron emission tomography (PET) biomarkers. Here, we aimed to investigate i) the diagnostic performance of blood biomarkers in the absence of CSF/PET modalities; ii) how these markers change longitudinally and relate to cognitive decline. Methods We evaluated cross‐sectional data for 307 participants (100 cognitively unimpaired (CU, Clinical dementia rating (CDR) score = 0) elderly and 207 cognitively impaired (CI; 101 with mild cognitive impairment (MCI) (CDR=0.5) and 106 AD dementia (CDR>0.5)) from the European multicentre AddNeuroMed cohort. 237 participants (76 CU, 161 CI) had longitudinal plasma biomarker and cognitive data (MMSE). Single molecule array (Simoa) assays were performed for NfL, t‐tau, Aβ42, Aβ40 and GFAp (all time‐points, single batch). P‐tau181 was measured using an in‐house Simoa assay. Cross‐sectional analyses were conducted using a generalized linear model for age‐adjustment, an ANOVA to compare biomarkers across groups, and correlation analysis to examine biomarker concentrations in relation to MMSE. Longitudinal analyses were performed using linear mixed models. Results At baseline, neuronal and astrocytic biomarkers were increased (p‐tau181, NfL, t‐tau, GFAp; P <0.001), whilst the amyloid marker (Aβ42/Aβ40 ratio) was decreased ( P <0.001) in AD compared with CU. Significant differences were also observed between AD and MCI for p‐tau181, Aβ42/Aβ40, GFAp ( P <0.001) and NfL ( P <0.05). P‐tau181 ( P <0.001) and NfL ( P <0.05) were increased in MCI compared with CU. ROC curves revealed high accuracy for p‐tau181 in separating AD from CU (AUC=0.921) and MCI (AUC=0.816), and in separating CU from CI (AUC=0.843). Linear mixed models showed increases in all five measures in both CU and CI subjects; when assessing change in MMSE, the strongest associations were seen for GFAp in CU (standardized‐β=‐0.27, P<0.001) and for p‐tau181 in CI subjects (standardized‐β=‐0.34, P<0.001). Conclusion These results suggest that blood biomarkers have diagnostic potential for AD in the absence of AT(N) classification, and that longitudinally, levels of p‐tau181 have potential as a biomarker of cognitive decline.