Causal associations between serum amino acid levels and osteoarthritis: a Mendelian randomization study.

ABSTRACT

OBJECTIVE: The association between serum amino acid (AA) levels and osteoarthritis (OA) risk remains unclear. METHOD: We performed a two-sample Mendelian randomization (MR) analysis to analyze the causal effects of serum AA levels on the OA risk by using summary-level genome-wide association study (GWAS) data. Inverse variance weighted (IVW) and Wald ratio were used as the main analysis. We also applied MR-Egger, Weighted median and Robust Adjusted Profile Score (MR.RAPS) methods. Heterogeneity and horizontally pleiotropic outliers were checked. The causal effects of AAs on early-onset all OA were explored. We also performed multivariable MR (MVMR) and conducted the bidirectional MR. RESULTS: The results suggested that genetically predicted alanine (Ala), tyrosine (Tyr) and isoleucine (Ile) levels were significantly associated with OA risk [e.g., association between Ala and hip/knee OA risk OR = 0.82, 95% confidence interval (CI) = 0.75-0.90, P = 1.54E-05]. The study yielded little evidence of associations between genetically predicted AA levels with early-onset all OA risk. When adjusting the body mass index (BMI) in the MVMR model, suggestive causal effects of Ala and Tyr were also identified, while the effects of Ile substantially attenuated with OA risk. No significant associations between OA and AA levels were observed after testing for bidirectionality. CONCLUSIONS: Some AAs, such as Ala, Tyr and Ile likely affects the OA risk especially at hip or knee joints. The findings highlight the important role that serum AAs might play in the development of OA and provided new treatment approaches to OA.