Plasma pTau181 and amyloid markers predict conversion to dementia in idiopathic REM sleep behaviour disorder

Abstract

Blood-based biomarkers for Alzheimers disease (AD) pathology have been investigated intensively as markers for AD-related neurodegeneration. Comorbid AD pathology is common in dementia with Lewy bodies (DLB). Accordingly, we hypothesized that plasma biomarkers associated with AD pathology might be useful to predict DLB in a cohort of idiopathic/isolated REM sleep behaviour disorder (iRBD), an incipient synucleinopathy. The aim of this study was to determine whether plasma amyloid- and pTau181 biomarkers can predict DLB. This longitudinal single-centre (Canada) cohort study included 158 individuals with polysomnography-confirmed iRBD between September 2004 and October 2022, each providing blood plasma samples, who were then offered prospective follow-up. Plasma A 40, A 42 and pTau181 levels were measured using NeuroToolKit, a prototype assay panel of neurodegeneration (Roche Diagnostics International Ltd). The primary outcome was the association between plasma biomarkers at baseline and eventual development of DLB. Correlations between plasma markers and baseline cognitive tests were assessed. A total of 142 iRBD participants [109 male (77%); mean "SD, 67.6 "8.0 years of age] were included in the final analysis. On prospective follow-up (2.9 "2.1 years after sampling), 32 individuals phenoconverted to a defined neurodegenerative syndrome (18 DLB, 13 Parkinsons disease and one multiple system atrophy). The combined phenoconvertor group had lower baseline plasma A 42/40 ratio compared with non-phenoconvertors (mean "SD, 0.103 "0.010 versus 0.114 "0.012, P < 0.001) and higher pTau181 levels (0.993 "0.354 versus 0.784 "0.266 pg/ml, P = 0.008). When divided by phenoconversion subtype, significant differences were seen selectively in DLB convertors [A 42/40 = 0.101 "0.010, difference -0.011, 95% confidence interval (-0.016; -0.005), P < 0.001; pTau181 = 1.144 "0.326 pg/ml, difference 0.282 pg/ml, 95% confidence interval (0.146; 0.418), P < 0.001]. Cross-sectional analysis showed that plasma pTau181 (but not A 42/40) was correlated with cognitive tests across various domains. Our results indicate that plasma A 42/40 ratio and pTau181 can predict conversion to DLB in iRBD.