Abstract: BACKGROUND: Osteonecrosis is a disabling and refractory disease with a high prevalence rate in China, necessitating the exploration of potential biomarkers for early prevention, diagnosis, and treatment. Metabolomic studies have demonstrated correlations between human metabolites and osteonecrosis; however, the causal relationship between plasma metabolites and osteonecrosis remains unclear. 
OBJECTIVE: To investigate the causal association between 1 400 plasma metabolites and osteonecrosis using Mendelian randomization and provide supporting evidence.
METHODS: Public data on 1 400 plasma metabolites (exposure factors) and osteonecrosis (outcome factor) were collected. The plasma metabolite data were derived from a genome-wide association study (GWAS) on blood metabolites published in Nature Genetics in January 2023, which included 1 091 blood metabolites and 309 metabolite ratios from 8 299 individuals in the Canadian Longitudinal Study on Aging (CLSA) cohort. The single-nucleotide polymorphism (SNP) data for osteonecrosis were obtained from the FinnGen R12 public database, comprising 475 307 samples (2 043 osteonecrosis cases and 473 264 controls). All of the participants were European ancestry. Mendelian randomization analysis was performed using RStudio (inverse-variance weighted, MR-Egger, weighted median, simple mode, and weighted mode methods), followed by heterogeneity testing, horizontal pleiotropy assessment, and Steiger directionality testing to ensure robustness and reliability. 
RESULTS AND CONCLUSION: (1) Three plasma metabolites were identified to have significant causal relationships with osteonecrosis (P < 0.05): Adenosine monophosphate-to-valine ratio (odds ratio=1.303, 95% confidence interval = 1.110-1.531, P=0.001,PFDR=0.07); oxidized cysteinylglycine levels (odds ratio=0.888, 95% confidence interval=0.791-0.998, P=0.046, PFDR=0.05); 3β,17β-androstenediol disulfate levels (odds ratio=1.121, 95% confidence interval=1.020-1.231, P=0.018, PFDR=0.06). (2) The adenosine monophosphate-to-valine ratio and 3β,17β-androstenediol disulfate levels were identified as risk factors for osteonecrosis, whereas oxidized cysteinylglycine levels acted as a protective factor. (3) These findings suggest a causal relationship between these three plasma metabolites and osteonecrosis, indicating their potential as biomarkers for early diagnosis and therapeutic targets for disease intervention. Although this study was based on European population data, it may provide valuable insights for osteonecrosis research in China. Future investigations by Chinese medical researchers may focus on detecting and modulating metabolite levels to achieve early diagnosis and precision treatment of osteonecrosis.

Key words: osteonecrosis, plasma metabolites, Mendelian randomization, genetics, causal relationship, single nucleotide polymorphism, genome-wide association study