Phenome-wide association of APOE alleles in the All of Us Research Program

Abstract

Background: Apolipoprotein E (APOE) variation is associated with altered lipid metabolism, as well as cardiovascular and neurodegenerative disease. Prior studies are largely limited to European ancestry populations and differential risk by sex and ancestry has not been widely evaluated.

Methods: We utilised a phenome-wide association study (PheWAS) to explore APOE-associated phenotypes in the All of Us Research Program. We determined APOE alleles for 181,880 participants, representing seven ancestry groups. We tested association of APOE variants, ordered based on Alzheimer's disease risk hierarchy (ε2/ε2 < ε2/ε3 < ε3/ε3 < ε2/ε4 < ε3/ε4 < ε4/ε4), with 2318 phenotypes. Bonferroni-adjusted analyses were performed overall, by ancestry, by sex, and with adjustment for social determinants of health (SDOH).

Findings: In the overall cohort, PheWAS identified 17 significant associations, including increased odds of hyperlipidaemia (OR 1.15 [1.14-1.16] per APOE genotype group; P = 1.8 × 10^-129), dementia, and Alzheimer's disease (OR 1.55 [1.40-1.70]; P = 5 × 10^-19), and reduced odds of fatty liver disease and chronic liver disease. ORs were similar after SDOH adjustment and by sex, except for an increased number of cardiovascular associations in males, and decreased odds of noninflammatory disorders of vulva and perineum in females. Significant heterogeneity was observed for hyperlipidaemia and mild cognitive impairment across ancestry. Unique associations by ancestry included transient retinal arterial occlusion in the European ancestry group, and first-degree atrioventricular block in the American Admixed/Latino ancestry group.

Interpretation: We replicate extensive phenotypic associations with APOE alleles in a large, diverse cohort. We provide a comprehensive catalogue of APOE-associated phenotypes and evidence of unique phenotypic associations by sex and ancestry, as well as heterogeneity in effect size across ancestry.

Funding: Funding is listed in the Acknowledgements section.

Keywords: Alzheimer’s disease; Apolipoprotein E; Health disparities; Hypercholesterolaemia; Phenome-wide association study.

Fig. 1

APOE PheWAS results from the entire cohort and its accompanying QQ plot. APOE genotypes were arranged hierarchically from ε2 to ε4 (ε2/ε2 < ε2/ε3 < ε3/ε3 < ε2/ε4 < ε3/ε4 < ε4/ε4). PheWAS analysis was adjusted for EHR length, age, sex, and PCs 1–16. In this cohort, 2318 phecodes with more than 50 cases or controls were analysed for association with APOE variants. The green horizontal line represents Bonferroni correction level (α = 2.16 × 10⁻⁵) and the red line represents the nominal significance (α = 0.05). Furthermore, the direction of arrows indicates either an increased or decreased OR. The x-axis represents the tested phecodes and the broader 18 categories they belong to, and the y-axis shows the −log₁₀(P-value).

Fig. 2

Volcano plot of P-values on y-axis against their corresponding odds ratios on x-axis after the PheWAS of the whole cohort. The 17 top hits in this analysis are tagged and labelled in orange. The plot shows AD had the largest effect size while hyperlipidaemia had the lowest P-value. Moreover, 14 associations are positively associated with APOE genotypes, whereas 3 phecodes are inversely associated. N.B. the y-axis is log transformed (log(−log₁₀(P-value))) for better visualisation with a cut-off at −2.

Fig. 3

Forest plots displaying the risk of disease and APOE diplotypes for a) hyperlipidaemia (Endocrine/Metabolic) b) dementias (Neurological) c) fatty liver disease (FLD) (Gastrointestinal) and d) coronary atherosclerosis [Atherosclerotic heart disease] (Cardiovascular). These plots illustrate the odds ratio and 95% CI after logistic regression analysis for each genotype compared to ε3/ε3, the reference group. Adjustment variables in the model included age at last event, sex, electronic health records length, high school education, median income, social deprivation index, and principal components 1–16. The phecodes represent the top signal within each phecode category from 17 top hits from the PheWAS for the whole cohort.

Fig. 4

Plot illustrates odds ratios from three separate analyses combined. The top hits from the PheWAS for the whole cohort were used to extract the similar phecodes from PheWAS summary outcomes in females and males. We can see that confidence intervals follow the similar patterns in three analyses and are larger in AD and dementias compared to other phecode categories. The results of heterogeneity test, I² and P-value, for each phecode is presented in the label of each corresponding phecode. One phecode, noninflammatory disorders of vulva and perineum, was excluded from the heterogeneity test because it was available for females. Four out of 16 tests indicated heterogeneity in results (P-values < 0.05). These phecodes were hyperlipidaemia, hypercholesterolaemia, pure hypercholesterolaemia, and mixed hyperlipidaemia.

Fig. 5

The forest plot displays the combined odds ratios from six separate analyses across ancestral groups. The 17 top hits from the PheWAS for the whole cohort were used to extract the similar phecodes from PheWAS summary outcomes in ancestral groups. Not all the above phecodes had case counts above 50 in all ancestral categories, such as AD appearing only for EUR and AMR. The results of heterogeneity test, I² and P-value, for each phecode is presented in the label of the corresponding phecode. Three associations including hyperlipidaemia, mild cognitive impairment, and symptoms and signs involving cognitive functions and awareness indicated heterogeneity in results (P-values < 0.05).