Pre-diagnostic plasma metabolomics and the risk of exfoliation glaucoma

ObjectiveTo identify pre-diagnostic plasma metabolomic biomarkers associated with risk of exfoliation glaucoma (XFG). MethodsWe conducted a metabolomic study using a 1:1 matched nested case-control study design within the Nurses Health Study (NHS) and Health Professionals Follow-up Study (HPFS). Participants provided blood samples in 1989-90 (NHS) and 1993-95 (HPFS); we identified 205 participants who newly developed XFG during follow-up to 2018 (average time to diagnosis from blood draw =11.8 years); XFG was confirmed with medical record review. We profiled plasma metabolites using liquid chromatography- mass spectrometry and identified 379 known metabolites that passed quality control checks. Metabolites were transformed using probit scores for normality. We used multivariable-adjusted logistic regression adjusting for matching factors (such as age, residential latitude, season and time of blood draw), glaucoma family history and other covariates. Metabolite Set Enrichment Analysis was used to identify metabolite classes associated with risk of XFG. Number of effective tests (NEF) and False Discovery Rate (FDR) were used to adjust for multiple comparisons. ResultsMean age of cases (n=205) at diagnosis was 71 years; 84% were women and >99% were Caucasian; matched controls (n=205) all reported eye exams as of the matched cases index date. A total of 33 metabolites were nominally significantly associated with XFG risk (p<0.05) and 4 metabolite classes were significantly associated (FDR<0.05). Overall, adverse associations were observed for the classes of lysophosphatidylcholines (FDR=0.02) and phosphatidylethanolamine plasmalogens (FDR=0.004). Inverse associations were observed for triglycerides (FDR<0.001) and steroid and steroid derivatives (FDR=0.03); in particular, the multivariable-adjusted odds ratio for XFG risk associated with each 1 standard deviation increase in plasma cortisone levels was 0.49 (95% CI=0.32-0.74; NEF=0.05). Results did not differ materially by time between blood draw and diagnosis, latitude of residence (< or [≥]41{degrees}N latitude), age (< or [≥]60 years), sex or glaucoma family history. ConclusionsFour broad classes of metabolites (including steroids such as cortisone and 3 lipid classes) in pre-diagnostic plasma collected almost a decade before diagnosis were associated with XFG risk; these results should be confirmed in future studies.