David D. Ward

Background and purpose The aim was to determine whether frailty is associated with the relationship between neuropsychological markers and global cognition in older adults. Methods Cross‐sectional analyzes were conducted of baseline data from three large cohort studies: National Alzheimer's Coordinating Center (NACC), Rush Memory and Aging Project (MAP) and Alzheimer's Disease Neuroimaging Initiative (ADNI). Studies recruited North American participants along the spectrum of cognitive functioning (44% no cognitive impairment at baseline). A frailty index was computed in each dataset. Frailty indices, neuropsychological tests (including measures of processing speed, episodic, semantic and working memory) and Mini‐Mental State Examination (MMSE) scores were the variables of interest, with age, sex, education and apolipoprotein E ε4 evaluated as confounders. Results Across all studies, 23,819 participants aged 55–104 (57% female) were included in analyzes. Frailty index scores were significantly and inversely associated with MMSE scores and significantly moderated relationships between neuropsychological test scores and MMSE scores. In participants with higher frailty index scores, lower neuropsychological test scores were more strongly associated with lower MMSE scores (standardized interaction coefficients ranged from −0.19 to −1.17 in NACC, −0.03 to −2.27 in MAP and −0.04 to −0.38 in ADNI, depending on the neuropsychological test). These associations were consistent across the different databases and were mostly independent of the composition of frailty indices (i.e., after excluding possible symptoms of dementia). Conclusions Amongst older Americans, frailty is associated with the cognitive expression of neuropsychological deficits. Implementation of frailty assessment in routine neurological and neuropsychological practice should be considered to optimize care outcomes for older adults.