**Background:** Intraocular pressure (IOP) is a major risk factor for glaucoma, a disease characterized by retinal ganglion cell loss. The mechanisms underlying IOP dysregulation and retinal ganglion cell homeostasis are poorly understood. Previous metabolomic studies have yielded inconsistent results regarding the relationship between plasma metabolites, IOP, and inner retinal thickness. This study aimed to systematically assess associations between 168 non-fasting plasma metabolites measured by nuclear magnetic resonance (NMR) spectroscopy and three key ophthalmic biomarkers: IOP, macular retinal nerve fiber layer thickness (mRNFL), and macular ganglion cell inner plexiform layer thickness (mGCIPL) in a large, well-characterized cohort.
**Methods:** This cross-sectional analysis used data from the UK Biobank, including 28,195 participants with IOP measurements (corneal-compensated IOPcc via Ocular Response Analyzer), 10,584 with mRNFL, and 10,554 with mGCIPL (measured by spectral-domain optical coherence tomography). Plasma metabolites were quantified using the Nightingale NMR platform. Multiple linear regression models were adjusted for demographics, lifestyle factors, medical history, and medication use (model 3). For inner retinal outcomes, additional adjustment for IOP was performed. Multiple comparisons were addressed using the number of effective tests (NEF) for individual metabolites and false discovery rate (FDR) for metabolite class analyses (Metabolite Set Enrichment Analysis, MSEA).
**Key Results:** In individual metabolite analysis, branched-chain amino acids (BCAAs) were significantly associated with lower IOP (total BCAAs: −0.12 mm Hg; 95% CI = −0.16 to −0.07; NEF = 2.7E-05). Other amino acids (phenylalanine, tyrosine, glutamine) also showed inverse associations. Conversely, albumin (0.16 mm Hg; 95% CI = 0.11 to 0.20; NEF = 9.6E-12), 3-hydroxybutyrate (0.18 mm Hg; 95% CI = 0.13 to 0.22; NEF = 8.9E-14), lactate (0.14 mm Hg; 95% CI = 0.09 to 0.18; NEF = 4.9E-08), and several lipids (including HDL and LDL cholesterol) were associated with higher IOP. For inner retinal outcomes, five HDL-related metabolites (e.g., phospholipids in medium HDL: −0.15 microns; NEF = 0.011) were associated with thinner mRNFL, while five LDL-related metabolites (e.g., free cholesterol in small LDL: 0.19 microns; NEF = 0.012) were associated with thicker mGCIPL. No metabolites were associated with thicker mRNFL or thinner mGCIPL. In MSEA, amino acids were the only metabolite class significantly associated with lower IOP (NES = −3.1 mm Hg; FDR = 1.6E-08). Lipoprotein subclass analysis revealed that medium HDL was associated with thinner mRNFL (NES = −2.2 microns; FDR = 1.8E-05) and mGCIPL (NES = −2.6 microns; FDR = 2.3E-07), while small LDL was associated with thicker mRNFL (NES = +1.9 microns; FDR = 0.0098) and mGCIPL (NES = +2.1 microns; FDR = 5.6E-06).
**Clinical Implications:** These findings suggest that plasma metabolites, particularly BCAAs and lipoproteins, may play distinct roles in IOP regulation and retinal ganglion cell integrity. The inverse association of BCAAs with IOP, though modest, warrants further investigation into their potential role in aqueous humor dynamics. The opposing associations of HDL (thinner mRNFL) and LDL (thicker mGCIPL) with inner retinal thickness support a model where LDL delivers lipids to the retina and HDL facilitates lipid efflux, with implications for retinal health and glaucoma risk. The results highlight the importance of lipid metabolism in retinal homeostasis and suggest that targeting specific lipoprotein pathways could offer new avenues for glaucoma prevention or therapy. However, the cross-sectional design and small effect sizes limit causal inference, and prospective studies are needed to confirm these associations.