Skip to main content
. 2021 Jun 23;12:660212. doi: 10.3389/fneur.2021.660212

Table 5.

Associations between polygenic risk score for cholesterol and cognitive status, in the Health and Retirement Study, European ancestry sample (n = 8,781)a.

High density lipoprotein cholesterol (HDL) Total cholesterol (TC)
CIND vs. normal Dementia vs. normal CIND vs. normal Dementia vs. normal
N OR (95 CI%) N OR (95 CI%) N OR (95 CI%) N OR (95 CI%)
Total effect of cholesterol PGS
Crude 7,707 1.03 (0.96, 1.10) 6,926 0.83 (0.71, 0.97) 8,590 0.98 (0.92, 1.04) 7,716 0.92 (0.79, 1.06)
Adjusted (demographic)b 7,707 1.02 (0.95, 1.10) 6,926 0.82 (0.69, 0.96) 8,590 0.97 (0.90, 1.05) 7,716 0.90 (0.76, 1.07)
Adjusted (health status)c 7,707 1.03 (0.96, 1.11) 6,926 0.83 (0.70, 0.98) 8,590 0.97 (0.90, 1.05) 7,716 0.90 (0.76, 1.07)
Direct effect of cholesterol PGS (adjusting for cholesterol level)
Crude 7,707 1.05 (0.98, 1.12) 6,926 0.84 (0.72, 0.98) 8,590 0.98 (0.92, 1.05) 7,716 0.92 (0.79, 1.06)
Adjusted (demographic) 7,707 1.03 (0.96, 1.11) 6,926 0.81 (0.69, 0.96) 8,590 0.97 (0.90, 1.04) 7,716 0.90 (0.76, 1.07)
Adjusted (health status) 7,707 1.04 (0.97, 1.12) 6,926 0.83 (0.70, 0.98) 8,590 0.97 (0.90, 1.05) 7,716 0.91 (0.76, 1.07)

APOE, Apolipoprotein E; CI, Confidence Interval; CIND, Cognitive Impairment-Non Dementia; OR, Odds Ratio.

a

All variables were measured at the first instance of biomarker collection for a participant from Health and Retirement Study waves 2006–2012. All the values were based on results from multivariable logistic regression analyses in each sample, in which “normal cognitive status” and “normal cholesterol level” were used as the reference groups.

b

Adjusted for age, sex, years of education, lipid-lowering medication, cholesterol measurement wave, and five ancestry-specific principal components.

c

Adjusted for ever drink alcohol, history of stroke, hypertension, diabetes, and BMI in addition to variables in b.