4,068 posters, 219 videos, 101 topics, 69 sessions, 14,365 authors, 6,128 institutions

Alzheimer's Association International Conference 2026

July 12-15, 2026 | London, United Kingdom, and Online

2463

Subjective Cognitive Decline-plus Features And Risk Of Conversion In A Community-based Cohort With Up To 16 Years Of Follow-up

View Poster Contact Presenter

Poster Presenter

Poosanu Thanapornsangsuth

Authors

Poosanu Thanapornsangsuth

Part of Topic

Subjective cognitive decline-plus features and risk of conversion in a community-based cohort with up to 16 years of follow-up

Poosanu Thanapornsangsuth1,2, Coty Chen3, Valentina Diaz3, Carina Lo3, Lana Callies3, Jessica Buxton3, Sophia Licata3, Shaista Afzal3, Albert Pham3, Laura Fenton3, Shubir Dutt3, Adam Staffaroni3, Emily W. Paolillo3, Rowan Saloner3, Kaitlin B. Casaletto2,3, Joel H. Kramer2,3.

Background

•SCD is considered an early stage preceding objective cognitive impairment.

•SCD-plus criteria aim to identify individuals at higher risk of progression.

•Previous studies mainly involved memory clinics or high-risk cohorts with short follow-up.

Aim Evaluate which SCD-plus features predict progression to MCI or dementia in a low-risk, community-based cohort followed for up to 16 years.

Methods

409 cognitively normal aged ≥50 years Brain Aging Network for Cognitive Health (BrANCH) cohort(2009–2023) (Table)

Predictor:

•39-item Everyday Cognition (ECog) questionnaire

•Self- and informant-reported concerns

Outcome: Consensus diagnosis of MCI or dementia

Analysis:

•Cox proportional hazards models

•Adjusted for age, sex, education, MMSE, and Geriatric Depression Scale scores

•Sensitivity analysis including APOE genotype

Results

•Baseline age was the strongest predictor of conversion.

•Age ≥65 years predicted progression better than the traditional ≥60-year SCD-plus cutoff.

•Any consistent cognitive symptom (ECog ≥3) or reported symptom (ECog ≥2) was more predictive than average ECog scores.

•Only the self-reported memory domain score predicted future impairment.

•Self-reported cognitive concernpredicted conversion, informant concern did not.

•Persistent self-reported global and attention complaintspredicted conversion.

•Persistent informant-reportedcomplaints were not predictive.

•APOE genotype was not associated with impairment-free survival.

Discussion

•Only selected SCD-plus featurespredicted long-term cognitive decline in this low-risk community cohort.

•Self-reported symptoms were consistently more informative than informant reports.

•SCD-plus criteria may require population-specific refinementrather than universal application.

•Findings can inform screening strategies and preclinical intervention trials.

•Harmonization of SCD-plus should consider both the target population and the feasibility of measuring individual features.

•

Contact

POOSANU THANAPORNSANGSUTH, M.D.

poosanu.thanapornsangsuth@gbhi.org

poosanu.t@chula.ac.th

Affiliations

1 King Chulalongkorn Memorial Hospital, Bangkok, Thailand, 2 Global Brain Health Institute, University of California, San Francisco, San Francisco, CA, USA, 3Edward and Pearl Fein Memory and Aging Center, University of California, San Francisco, San Francisco, CA, USA.

Figure Results of Cox proportional hazards models. Hazard ratios are shown for baseline ECog scores reported by the participant (self), informant, or both. In Panel A, the left plot displays hazard ratios per standard deviation increase in average domain scores; for the combined measure, scores represent the sum of self- and informant-reported ECog. The middle and right plots in Panel A show hazard ratios for the presence of any item within a domain scored as consistent (≥3) or reported (≥2), respectively. Panel B shows hazard ratios for the presence of cognitive concern reported by the participant and/or the informant (N refers to all participants withavailable data). In participants with longitudinal ECog data, Panel C presents hazard ratios for persistent ECogsymptoms, defined as individuals whose slope and intercept from linear mixed-effects models were at or above the 50th percentile.

Comments

Email is required but is not shown publicly.

Sort comments

Loading comments...