Subjective memory decline and neuropsychological performance: Epidemiology / Risk and protective factors in MCI and dementia
Ying Xu,Ruth Peters,Ranmalee Eramudugolla,Kaarin J Anstey
DOI: https://doi.org/10.1002/ALZ.045656
2020-01-01
Alzheimer s & Dementia
Abstract:Background The association between subjective memory decline (SMD) and different domains of cognition is not well understood. Method Four waves (4‐year intervals between waves) of data from 2,551 participants (aged 62.5 ±1.5 years, 52% male) of the Personality and Total Health (PATH) study were used. SMD was defined as a response of ‘no’ to one question: ‘Do you feel you can remember things as well as you used to?’ (w1 and w2), and categorized as ‘consistent’ (SMD both waves), ‘fluctuate’ (either wave) and ‘neither’. The more comprehensive Memory Assessment Clinic‐Q (MAC‐Q) was also available at w4, and includes remembering names, phone numbers, postcodes, where objects are, newspaper reports and shopping‐list items. Neuropsychological tests (w2 to w4) included immediate and delayed recall, digit span backwards, spot the word, symbol digit modalities (SMDT), Trails A and B and Purdue pegboard. We used random‐effect linear mixed models (LMMs) to examine temporal and prospective associations between SMD categories and neuropsychological performance, and linear regression models to investigate cross‐sectional associations between SMD and neuropsychological performance at w4. Result Fifty‐one percent, 27% and 22% of participants had ‘neither’, ‘fluctuate’ and ‘consistent’ SMD. SMD categories were not associated with steeper decline in any neuropsychological performance, but were associated with lower z transformed SDMT and purdue pegboard (both hands and dominant hand) scores at w2 in all LMMs (adjusted for age, sex, education [model 1], plus anxiety, depression [model 2], smoking, hypertension, diabetes and high cholesterol [model 3], and effect sizes were ‐0.07 (‐0.13, ‐0.01), ‐0.20 (‐0.32, ‐0.08) and ‐0.13 (‐0.24, ‐0.02) in model 3 (all P ≤ 0.03). In sensitivity analyses, similar results were found when using SMD at w1 or w2, separately, as the ‘exposure’. Nineteen percent of participants had MAC‐Q assessed SMD, which was cross‐sectionally associated with all neuropsychological tests except for digit span backwards (all P<0.02). Conclusion Presence of MAC‐Q assessed SMD was cross‐sectionally associated with verbal memory, executive function, processing speed, manual dexterity and bimanual coordination. To be useful in healthcare settings, while ensuring simplicity, SMD measurement may need to contain specific memory items, and its value in predicting cognitive decline needs confirmation.