Predicting progression from subjective cognitive decline to dementia using different neuropsychological criteria: a longitudinal study

Background

Subjective Cognitive Decline (SCD) is increasingly recognized as a preclinical stage associated with an elevated risk of future cognitive impairment and dementia. It is conceptualized as a self-experienced decline in cognitive capacity in individuals who perform within normal limits on standardized neuropsychological testing. However, a major challenge in the field lies in the absence of consensus regarding the operational definition of “normal” cognitive performance. Different neuropsychological cut-offs for Mild Cognitive Impairment (MCI), particularly those based on varying stringency levels, may substantially influence the classification of individuals as cognitively normal and, consequently, the identification of SCD. This study aimed to evaluate whether distinct neuropsychological criteria used to define cognitive normality impact the prognostic validity of SCD in predicting conversion to dementia.

Methods

Participants were drawn from the Cognitive Complaints Cohort (CCC) and diagnosed with SCD according to the Subjective Cognitive Decline Initiative (SCD-I) criteria. Cognitive normality was operationalized as the absence of MCI based on five alternative neuropsychological approaches derived from the Jak and Bondi framework, ranging from more liberal to more conservative definitions. Descriptive statistics were used to characterize sociodemographic, clinical, and neuropsychological variables across SCD groups. A bootstrap resampling procedure was applied to estimate means and 95% confidence intervals (CI) for key parameters, including age, education, follow-up time, conversion to dementia, Subjective Memory Complaints (SMC) scale, Mini-Mental State Examination (MMSE), Blessed Dementia Rating Scale – first part (BDRS), and Geriatric Depression Scale (GDS). Time-to-event analyses were conducted using Kaplan–Meier survival curves to estimate dementia incidence across SCD definitions, and differences between groups were assessed using the log-rank test.

Results

A total of 832 participants were included. The application of different neuropsychological criteria resulted in SCD groups that were broadly comparable in age and sex distribution but differed significantly in educational attainment and baseline cognitive and functional profiles. The Conservative SCD group exhibited lower educational levels (mean 10.1 years, 95% CI 9.8–10.4), lower global cognitive performance (MMSE mean 26.9, 95% CI 26.7–27.1), and greater functional impairment (BDRS mean 0.8, 95% CI 0.6–0.9) compared to the Liberal SCD group (education mean 11.4 years, MMSE mean 28.1, BDRS mean 0.3). No significant differences were observed across groups in subjective memory complaints or depressive symptoms. Kaplan–Meier analyses demonstrated marked variability in dementia risk according to SCD definition. Log-rank tests indicated significantly higher and faster conversion rates in the Conservative, Comprehensive, and Historical SCD groups compared with the Liberal and Typical groups (p < .005). Five-year conversion probabilities ranged from 3.9% (Liberal SCD) to 25.5% (Conservative SCD), increasing to 16.2% and 40.9%, respectively, at 10 years.

Conclusions

The use of different neuropsychological criteria to define cognitive normality yields clinically and prognostically distinct SCD populations. These findings underscore the critical importance of explicitly specifying diagnostic thresholds in both research and clinical settings, as they substantially influence patient characterization and long-term risk stratification for dementia.

References

No references