Association between the metabolome and bone mineral density in a Chinese population

Background: Osteoporosis is a common metabolic bone disease, which always leads to osteoporotic fractures. Biomarkers of bone mineral density (BMD) are helpful for prevention and early diagnosis of osteoporosis. This study aims to identify metabolomic biomarkers of low BMD. Methods: We included 701 participants who had BMD measures by dual-energy X-ray absorptiometry scans and donated fasting plasma samples from three clinical centres as a discovery set and another 278 participants from the fourth centre as an independent replication set. We used a liquid chromatography-mass spectrometry-based metabolomics approach to profile the global metabolites of fasting plasma. Findings: Among the 265 named metabolites identified in our study, six were associated with low BMD (FDR-adjusted P<0.05) in the discovery set and were successfully validated in the independent replication set. The circulating levels of five metabolites, i.e., inosine, hypoxanthine, PC (0-18:0/22:6), SM (d18:1/21:0) and isoleucyl-proline were associated with decreased odds of low BMD, and PC (16:0/18:3) level was associated with increased odds of low BMD. Per 1-SD increase in a composite metabolite score of these six metabolites was associated with about half decreased odds of low BMD (odds ratio 0.59, 95% confidence interval: 0.52-0.68). Furthermore, introduction of a panel of metabolites selected by elastic net regression to a prediction model of classical risk factors and plasma biomarker of bone resorption substantially improved the prediction performance for low BMD (ADCs: 0.782 vs. 0.698, P=0.002). Interpretation: Metabolomics profiling may help identify novel biomarkers of low BMD and be helpful for early diagnosis of osteoporosis beyond the current clinical index. (C) 2020 The Authors. Published by Elsevier B.V.