Diagnostic and prognostic performance and longitudinal changes in plasma neurofilament light chain concentrations in adults with Down syndrome: a cohort study
The Lancet Neurology | August 1, 2021
Carmona-Iragui M, Alcolea D, Barroeta I, Videla L, Muñoz L, Van Pelt KL, Schmitt FA, Lightner DD, Koehl LM, Jicha G, Sacco S, Mircher C, Pape SE, Hithersay R, Clare ICH, Holland AJ, Nübling G, Levin J, Zaman SH, Strydom A, Rebillat AS, Head E, Blesa R, Lleó A and Fortea J
Lancet Neurol. 2021 Aug;20(8):605-614
DOI: https://doi.org/10.1016/S1474-4422(21)00129-0\
Abstract
Background: Adults with Down syndrome are at an ultra-high risk of Alzheimer’s disease, but diagnosis of Alzheimer’s disease in this population is challenging. We aimed to validate the clinical utility of plasma neurofilament light chain (NfL) for the diagnosis of symptomatic Alzheimer’s disease in Down syndrome, assess its prognostic value, and establish longitudinal changes in adults with Down syndrome.
Methods: We did a multicentre cohort study, including adults with Down syndrome (≥18 years), recruited from six hospitals and university medical centres in France, Germany, Spain, the UK, and the USA, who had been assessed, followed up, and provided at least two plasma samples. Participants were classified by local clinicians, who were masked to biomarker data, as asymptomatic (ie, no clinical suspicion of Alzheimer’s disease), prodromal Alzheimer’s disease, or Alzheimer’s disease dementia. We classified individuals who progressed along the Alzheimer’s disease continuum during follow-up as progressors. Plasma samples were analysed retrospectively; NfL concentrations were measured centrally using commercial kits for biomarker detection. We used ANOVA to evaluate differences in baseline NfL concentrations, Cox regression to study their prognostic value, and linear mixed models to estimate longitudinal changes. To account for potential confounders, we included age, sex, and intellectual disability as covariates in the analyses.
Findings: Between Aug 2, 2010, and July 16, 2019, we analysed 608 samples from 236 people with Down syndrome: 165 (70%) were asymptomatic, 32 (14%) had prodromal Alzheimer’s disease, and 29 (12%) had Alzheimer’s disease dementia; ten [4%] participants were excluded because their classification was uncertain. Mean follow-up was 3·6 years (SD 1·6, range 0·6-9·2). Baseline plasma NfL concentrations showed an area under the receiver operating characteristic curve of 0·83 (95% CI 0·76-0·91) in the prodromal group and 0·94 (0·90-0·97) in the dementia group for differentiating from participants who were asymptomatic. An increase of 1 pg/mL in baseline NfL concentrations was associated with a 1·04-fold risk of clinical progression (95% CI 1·01-1·07; p=0·0034). Plasma NfL concentrations showed an annual increase of 3·0% (95% CI 0·4-5·8) per year in the asymptomatic non-progressors group, 11·5% (4·9-18·5) per year in the asymptomatic progressors group, and 16·0% (8·4-24·0) per year in the prodromal Alzheimer’s disease progressors group. In participants with Alzheimer’s disease dementia, NfL concentrations increased by a mean of 24·3% (15·3-34·1).
Interpretation: Plasma NfL concentrations have excellent diagnostic and prognostic performance for symptomatic Alzheimer’s disease in Down syndrome. The longitudinal trajectory of plasma NfL supports its use as a theragnostic marker in clinical trials.