Lipoprotein(a) and Cognitive Function in the Northern Manhattan Study

Background

Elevated lipoprotein(a) [Lp(a)] is an established cardiovascular risk factor, but its relationship with cognitive performance and cognitive decline remains uncertain. We examined associations between Lp(a) concentrations and cognitive function in participants from the Northern Manhattan Study (NOMAS).

Methods

We analyzed 188 participants from the NOMAS MRI/neuropsychological subcohort with baseline Lp(a) measurements and repeated cognitive assessments over approximately 10 years. Lp(a) was evaluated continuously (per 10 mg/dL), dichotomized at >50 mg/dL, and as the highest decile (top 10%). Cross-sectional associations with global cognition, memory, language, executive function, and processing speed were examined using multivariable linear regression. Longitudinal cognitive trajectories were evaluated using linear mixed-effects models with participant-specific random intercepts. Model 1 adjusted for age, sex, race, and education; Model 2 additionally adjusted for hypertension, diabetes, and body mass index; and Model 3 further adjusted for smoking.

Results

The mean baseline Lp(a) concentration was 38.7 mg/dL, and 30.9% of participants hadLp(a) concentrations >50 mg/dL. Cross-sectionally, higher continuous Lp(a) was associated with better processing speed across all models; in the fully adjusted model, each 10 mg/dL higher Lp(a) was associated with higher processing speed (β=0.044, P=0.003). Participants in the highest Lp(a) decile also demonstrated better baseline processing speed than the remaining participants (β=0.450, P=0.026). No significant cross-sectional associations were observed for global cognition, memory, language, or executive function. Longitudinally, continuous Lp(a) and Lp(a) >50 mg/dL were not associated with cognitive decline in any cognitive domain. However, participants in the highest Lp(a) decile experienced a faster decline in processing speed over time (Lp(a)×time interaction β=−0.066, 95% CI −0.125 to −0.008, P=0.025), while no significant longitudinal associations were observed for memory, language, or executive function.

Conclusions

Overall, Lp(a) was not consistently associated with baseline cognitive performance or longitudinal cognitive decline across multiple cognitive domains. Although the observed association with processing speed should be interpreted cautiously it suggests a potential vascular pathway as processing speed is particularly sensitive to cerebrovascular injury. This finding may reflect the effects of Lp(a) on cerebral small vessel disease that requires further confirmation.