Predicting Alzheimer's Disease from Combined Slow Gait and Low Activity Fragmentation - is Activity Fragmentation Always Bad?

Abstract

Background: Among older people, slow walking is a mild predictor of Alzheimer's disease (AD). However, studies that assessed this association did not consider slow walking may have different causes and that the increased risk of AD may be limited to a subgroup. We hypothesized that individuals with slow walking who show frequent alternative bouts of activity and rest are more likely to have non-neurological causes of slow walking, while those who fail to use this compensatory strategy despite slow walking are more likely to have neurological causes that put them at higher risk of AD. Methods: We analyzed data on gait speed and activity fragmentation (AF) by accelerometry from 505 initially cognitively normal persons aged 60 and older (51%women, 24%Black). New diagnoses of mild cognitive impairment (MCI) or AD were adjudicated over a mean of 7·2-years. Associations of gait speed, AF, and their interaction with incident MCI/AD were evaluated by Cox proportional-hazards models, adjusted for covariates. Findings: MCI/AD developed in 64 participants. Each 0·05m/sec slower gait speed was associated with 8% higher risk of MCI/AD (p=0·03). AF alone was not associated with MCI/AD risk (p>0·05), but we found a significant “gait*AF” interaction (p=0·02). At low AF (-1SD), each 0·05 m/sec slower gait speed was associated with 16% higher risk of MCI/AD, while at higher AF (+1SD), gait speed did not predict MCI/AD. Among those with slow gait, higher AF was associated with less decline in pegboard dominant hand performance, higher odds of having lower extremity osteoarthritis, and higher cardiopulmonary burden. Interpretation: Frequent rests among slow walkers are associated with lower risk of future MCI/AD, perhaps because this behavioral strategy indicates lower likelihood of subclinical brain damage. On the contrary, failing to fragment activity in the face of impaired mobility may indicate subclinical brain damage and, therefore, is predictive of MCI/AD. Funding: This study was supported in part by the Intramural Research Program of the National Institute on Aging. Declaration of Interest: None declared. Ethical Approval: The BLSA protocol was approved by the Institutional Review Board of the National Institutes of Health. Participants provided written informed consent at each BLSA visit.