Distinct trajectories of subjective cognitive decline before diagnosis of neurocognitive disorders: Longitudinal modelling over 18 years.

Background: Subjective cognitive decline (SCD) is an established predictor of neurocognitive disorders (NCD) (i.e. mild cognitive impairment and dementia). Yet, its construct remains contentious. Many individuals with SCD do not progress to NCD, leading to an alternative term in the literature - 'functional cognitive disorders' - to describe the SCD experience in these individuals.

Objectives: To examine the distinct differences in trajectories of SCD between those who did and did not eventually develop NCD.

Design: Case-control study.

Setting: Alzheimer's Disease Centers across USA.

Participants: A total of 5,167 participants aged ≥50 years were followed up near-annually to evaluate for SCD and NCD (median follow-up=8.1 years; range=1.0-18.0). Cases were defined as those who developed incident NCD during follow-up; controls completed ≥10 years of follow-up and had normal cognition throughout follow-up period.

Measurements: SCD was evaluated with a yes/no question based on "perceived decline in memory relative to previously attained abilities". The trajectories of SCD were modelled with mixed-effect logistic regression, using a backward timescale.

Results: Those who developed NCD (cases) had new onset of SCD within past 20 years, which became particularly noticeable 13-14 years before diagnosis, and became even more evident in the last 4 years. Those who did not develop NCD (controls) reported SCD since younger age, with the probability of SCD remaining constant over time. The distinctive trajectories were consistent across Alzheimer's and non-Alzheimer's disease, and among those with higher baseline rates of SCD due to psychiatric conditions.

Conclusions: SCD exhibits distinctive trajectories among those who do and do not progress to NCD. These distinctive trajectories can inform NCD risk for early interventions, and guide public health messaging to distinguish high-risk SCD from normal ageing. Future SCD scales may possibly need to evaluate symptom changes over a longer, 20-year horizon to better capture the new onset of SCD within this longer timeframe.