Using the National Institutes of Health’s All of Us Research Program data, a new cross-sectional analysis revealed a significant U-shaped association between serum high-density lipoprotein (HDL) levels and risk for age-related macular degeneration (AMD).
The study, which was recently published in Ophthalmology, further implicated specific HDL-related genetic variants in AMD susceptibility and identified lipoprotein(a) (Lp[a]) as a novel single nucleotide polymorphism (SNP) that is associated with increased disease risk.
The analysis included 2,328 patients with AMD and 5,028 age-, sex-, and race-matched healthy controls. Clinical data were derived from laboratory values and medical histories, while genomic associations were assessed using SNP data that was processed via the PLINK toolkit. The research leveraged multivariable logistic regression models to identify associations between lipoprotein metrics and AMD status.
The researchers found that both low and high HDL levels were significantly associated with increased AMD risk (P<.001 for both). This relationship formed a U-shaped curve in an odds ratio analysis. The lowest AMD risk occurred around HDL levels between 40 and 60 mg/dL. Associations between HDL and AMD were independent of gender, smoking history, statin use, or presence of cardiovascular disease.
Low and high LDL and triglyceride levels were not significantly associated with AMD (P=.32–.56).
After incorporating genomic data into the analysis, the researchers found that CFH (rs1061170) and ARMS2 (rs10490924) SNPs were significantly associated with AMD (odds ratio [OR] = 2.04 and 1.64, respectively; both P<.001). ABCA1 (rs3890182) and LIPC (rs1800588), both involved in HDL metabolism, were inversely associated with AMD risk (ABCA1 OR = 0.88, P=.04; LIPC OR = 0.86, P=.001). LPA (rs3798220), which is associated with lipoprotein(a), was newly linked to increased AMD risk (OR = 1.37, P=.007).
These findings align with prior studies that implicated complement and lipid pathways in AMD and add new insight into HDL-mediated risk.
The authors propose that HDL may contribute to drusen formation—the hallmark of early and intermediate AMD—through local aggregation in Bruch’s membrane. Elevated systemic HDL might increase HDL deposition at the retina, leading to inflammation and oxidative stress. This hypothesis builds on prior immunohistochemical evidence that showed HDL and oxidized lipids in drusen.
At very high HDL levels, the risk of AMD in this cohort approached an odds ratio of 1.5, which paralleled observations in cardiovascular studies (OR = 1.6–4.5) for high HDL and increased AMD risk.
The retrospective nature of the study and reliance on ICD codes limited AMD severity classification. Subgroup analysis by disease stage was not performed due to missing severity data in 34% of AMD cases. Serum Lp(a) levels could not be assessed due to data sparsity.
The study authors detailed avenues for future research. “The precise mechanism for why elevated HDL contributes to [cardiovascular disease] and AMD is not completely understood,” they wrote. “One possible hypothesis is that higher serum concentrations contribute to higher local concentrations of HDL in Bruch's membrane, which result in retention and aggregation and serve as a nidus for oxidation and inflammation in AMD pathogenesis. More work is needed to understand how both low and high HDL levels play a role in AMD pathogenesis, which may guide our understanding of the impact of lipid dynamics both systemically and locally in the eye.”
They added that associations between Lp(a) and drusenogenesis may be possible as All of Us enrollment continues to expand.
“These findings highlight an exciting application of the wealth of clinical and genetic data in the All of Us data set to highlight novel associations with the retinal disease pathophysiology of AMD,” they concluded. RP