Scientific Reports | June 29, 2021

Pape SE, Al Janabi T, Ashton NJ, Hye A, Sheehan R, Gallagher P, Knight B, Prins AM, Courtenay K, Jordanova V, Thomas B, Perumal N, Forbes C, Hassiotis A and Strydom A

Sci Rep. 2021;11:13438

DOI: https://doi.org/10.1038/s41598-021-92887-5

Abstract

The validity of dementia diagnostic criteria depends on their ability to distinguish dementia symptoms from pre-existing cognitive impairments. The study aimed to assess inter-rater reliability and concurrent validity of DSM-5 criteria for neurocognitive disorder in Down syndrome. The utility of mild neurocognitive disorder as a distinct diagnostic category, and the association between clinical symptoms and neurodegenerative changes represented by the plasma biomarker neurofilament light were also examined. 165 adults with Down syndrome were included. Two clinicians independently applied clinical judgement, DSM-IV, ICD-10 and DSM-5 criteria for dementia (or neurocognitive disorder) to each case. Inter-rater reliability and concurrent validity were analysed using the kappa statistic. Plasma neurofilament light concentrations were measured for 55 participants as a marker of neurodegeneration and between group comparisons calculated. All diagnostic criteria showed good inter-rater reliability apart from mild neurocognitive disorder which was moderate (k = 0.494). DSM- 5 criteria had substantial concurrence with clinical judgement (k = 0.855). When compared to the no neurocognitive disorder group, average neurofilament light concentrations were higher in both the mild and major neurocognitive disorder groups. DSM-5 neurocognitive disorder criteria can be used reliably in a Down syndrome population and has higher concurrence with clinical judgement than the older DSM-IV and ICD-10 criteria. Whilst the inter-rater reliability of the mild neurocognitive disorder criteria was modest, it does appear to identify people in an early stage of dementia with underlying neurodegenerative changes, represented by higher average NfL levels.

This study was performed using the Quanterix HD-1 Analyzer.