Related Experiment Video
Updated: May 28, 2026
Dissection of Human Retina and RPE-Choroid for Proteomic Analysis
Published on: November 12, 2017
Plasma Metabolomic Associations with Incident Rhegmatogenous Retinal Detachment in UK Biobank
Yuelan Gao1, Jianming Xu1, Jun Yu1
1Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China.
Objective:
To investigate the associations of plasma metabolites with incident rhegmatogenous retinal detachment (RRD).
Design:
A prospective cohort study.
Subjects:
A total of 213 173 UK Biobank participants.
Methods:
Plasma metabolomics were quantified using the Nightingale Health platform based on proton nuclear magnetic resonance spectroscopy, covering 251 metabolites. Associations with incident RRD were assessed using Cox proportional hazard models and restricted cubic splines. Two-sample Mendelian randomization analyses were conducted to assess whether genetically predicted metabolite levels showed patterns consistent with the observational associations. A metabolomic risk score (MRS) was constructed based on significant metabolites to evaluate their combined effects on RRD. Principal component analysis was used to test whether the metabolite profiles improved RRD risk prediction.
Main Outcome Measures:
Incident RRD, defined as the first recorded diagnosis during follow-up.
Results:
Over a median follow-up of 11.10 years, 1111 participants developed RRD. Seventeen metabolites were significantly associated with RRD after false discovery rate correction (P FDR < 0.05), including 11 high-density lipoprotein (HDL)-associated metabolites, 4 fatty acids, 1 very low-density lipoprotein (VLDL)-associated metabolite, and creatinine. Among them, omega-3, docosahexaenoic acid (DHA), and 11 HDL-associated metabolites were inversely associated with RRD, whereas monounsaturated fatty acids to total fatty acids, creatinine, and free cholesterol to total lipids in large VLDL showed positive associations. Mendelian randomization analyses provided supportive genetic evidence consistent with the inverse associations for omega-3, DHA, phospholipids in very large HDL, and total lipids in large HDL. Compared with participants in the low MRS group, those with high MRS had a significantly higher risk (hazard ratio = 1.608, 95% confidence interval: 1.187-2.178, P = 0.002). However, the RRD-associated metabolites did not materially improve model discrimination (C-index 0.694-0.697 across models).
Conclusions:
This study identified 17 plasma metabolites associated with RRD risk, particularly omega-3 and HDL-related lipids. Although metabolomics did not improve risk prediction, these findings provide exploratory evidence of systemic metabolic patterns associated with RRD risk.
Financial Disclosures:
The authors have no proprietary or commercial interest in any materials discussed in this article.
