81 Markers of neurodegeneration in subjective cognitive decline Classification of evidence This study provides Class II evidence that HV, NfL and GFAP predicted clinical progression beyond A and T in individuals with SCD. INTRODUCTION In recent years, there has been a major change in the definition of Alzheimer’s disease (AD). Formerly, the core criteria of AD diagnosis were based on clinical symptoms (1). In 2018, a research framework has been put forward by the NIA-AA in which every individual is classified based on specific biomarkers in the AT(N) classification (2). In this framework, the term ‘Alzheimer’s disease’ refers to the presence of abnormal amyloid-beta accumulation and neurofibrillary tau tangles, i.e. ‘A’, measured by CSF Aβ or amyloid PET, and ‘T’, measured by CSF phosphorylated tau (p-tau) or tau PET. The AT(N) construct is independent of the cognitive stage of the individual, which makes it possible to identify AD in cognitively normal individuals. The ‘N’ in the AT(N) classification represents neurodegeneration. Neurodegeneration can have many different causes and is not specific for AD. Therefore, neurodegenerative markers are not necessary for the diagnosis, but rather have been suggested to provide pathologic staging information and predictive value. Proposed biomarkers of N include atrophy on MRI, hypometabolism on fluorodeoxyglucose (FDG) PET or CSF total tau (t-tau) (2). In addition, blood-based biomarkers are now available and have been suggested as non-invasive alternative markers for N (2-4). Allowing different biomarkers as indicator of a biomarker group implies that they can be used interchangeably and measure the same pathological process. For the A and T biomarker group, this assumption holds fairly well, with moderate to high agreement and relatively high correlation coefficients between markers within A and T, respectively (5-7). N biomarkers, however, are poorly correlated and show inadequate agreement (6, 8-11). Furthermore, the fact that N biomarkers are suggested to provide staging information implies that individuals with a higher degree of neurodegeneration are assumed to deteriorate faster. However, there are only a few studies that directly compared different N biomarkers in their association with clinical progression or cognitive decline over time. Most are hampered by small sample sizes, and none have directly compared blood-based biomarkers to CSF and imaging biomarkers yet (10, 12-15). It is difficult to determine which modality captures ‘neurodegeneration (N)’ most accurately, because there is no gold standard available. However, it should capture a different process than the accumulation of amyloid-beta (A) or fibrillary tau (T), as 4