Among individuals with a positive Alzheimer’s disease (AD) biomarker or an apolipoprotein E (APOE) ɛ4 allele, arterial stiffening reflected by higher pulse wave velocity (PWV) is associated with poorer cognition, particularly in episodic memory and executive function domains. However, it remains unclear whether arterial stiffening also predicts longitudinal decline in cognition in older adults. We therefore examined associations between PWV and AD biomarker positivity on longitudinal cognitive performance in well-characterized older adults drawn from the UC San Diego Shiley-Marcos Alzheimer’s Disease Research Center.

152 older adults without dementia (138 without cognitive impairment, 14 with mild cognitive impairment) from the UC San Diego Shiley-Marcos Alzheimer’s Disease Research Center (ADRC) underwent APOE genotyping, PWV assessment with a SphygmoCor system, and neuropsychological testing. Participants were included if they completed baseline neuropsychological testing and at least 1 follow up assessment within 3 years following baseline. Participants were considered APOE ɛ4-positive if they had 1 or 2 ɛ4 alleles (61 ɛ4-positive; 91 ɛ4-negative). Neuropsychological test scores were converted to z-scores based on data from a group of robust normal controls and averaged across 5 cognitive domains to create composite measures for memory, executive function, attention, language, and visuospatial function. CSF was collected by lumbar puncture within 3 years of pulse wave velocity measurement for a subset of participants (n=102) and analyzed for AD biomarkers. AD biomarker positivity was defined using a cutoff of CSF Aβ42/Aβ40 < 0.056 (n_{biomarker-neg}=72, n_{biomarker-pos}=30). Linear mixed effects models examined if PWV (1) predicted cognitive decline, and if (2) interacted with APOE ɛ4 status or CSF AD biomarker status to impact cognitive decline. Covariates included sex, education, age, cognitive diagnosis, APOE ɛ4 status and Framingham Stroke Risk Profile at baseline.

In general, higher baseline PWV was not significantly associated with accelerated decline in any cognitive domain. However, PWV interacted with AD biomarker positivity to predict decline in language (p=0.011), such that higher PWV predicted faster language decline in AD biomarker-positive individuals relative to biomarker-negative individuals. Additionally, PWV interacted with APOE-ɛ4 status to predict decline in visuospatial function (p=.023) such that higher PWV predicted faster visuospatial decline in ɛ4-negative individuals relative to ɛ4-positive individuals.

The interaction between vascular and AD-related contributions to cognitive decline varies by cognitive domain. In particular, arterial stiffing as measured by higher PWV predicts significantly faster language decline in AD biomarker-positive individuals relative to AD biomarker-negative individuals. Conversely, higher PWV is associated with greater visuospatial decline in ɛ4-negative individuals relative to ɛ4-positive individuals. Results underscore the importance of investigating arterial stiffening as it relates to cognitive decline in older adults. Findings also provide further support for the existence of a synergistic relationship between neurodegenerative markers and elevated vascular risk in at-risk older adults. PWV may be a useful measure for prediction of decline in various cognitive abilities in older adults at elevated risk of AD vis a vis genetic or biomarker status.