Abstract Background Metabolic dysfunction-associated steatotic liver disease (MASLD) is linked to cardiovascular risk, necessitating an exploration of potential metabolic contributors, such as amino acids. While amino acids are implicated in metabolic disorders, their specific relationship with MASLD and subsequent cardiovascular complications is not well-defined. We aimed to delineate the association between circulating amino acids and MASLD, identifying amino acids that could potentially link MASLD to cardiovascular issues. Methods Utilizing UK Biobank data, we tested the association of ten amino acids with MASLD in a cohort of 72,626 MASLD cases and 128,102 controls. Afterwards, we performed Mendelian Randomization (MR) using genome-wide association study (GWAS) data from 8,434 MASLD cases and 770,180 controls, and GWAS of amino acids including 115,052 participants from UK Biobank. Multiple MR methods, including MR-Egger, Cochran's Q-test, and MR-PRESSO, were applied to assess pleiotropy and heterogeneity. Survival analysis in MASLD patients assessed the link between significant amino acid and incident severe cardiovascular outcomes, including stroke, heart failure, myocardial infarction, and mortality, using three different covariate combinations: Model 1 included age, sex, and fasting time; Model 2 included in addition smoking, alcohol use, education, physical activity, and Model 3 included body mass index (BMI) in addition to model 2. Results Nine out of the ten amino acids were statistically significantly associated with MASLD across all three models (p<6.25×10^-3). Among them valine, leucine, isoleucine, total branched-chain amino acids, tyrosine, alanine, and phenylalanine were positively associated with MASLD, while glutamine and glycine were negatively associated with MASLD. The genetic predisposition of MASLD was significantly associated with increased phenylalanine (beta=0.05, p=4.0×10^-4) and tyrosine (beta=0.15, p=2.0×10^-3), suggesting that alterations in these amino acids may result from the progression of MASLD. We also detected heterogeneity (Cochran’s Q-test: p=8.4×10^-9) and pleiotropy (MR-Egger: p=0.02) for tyrosine, but not for phenylalanine. Furthermore, phenylalanine levels in MASLD patients were significantly (p<6.25×10^-3) associated with the incidence of heart failure, stroke, and mortality, but not with myocardial infarction across all three models. We did not identify a significant association between baseline tyrosine levels and any cardiovascular complications or mortality across the three models (Figure 1). Conclusion Our findings indicate a strong association between certain amino acids and MASLD, with phenylalanine, in particular, emerging as a potential mediator of cardiovascular risk in MASLD patients. The absence of heterogeneity and pleiotropy for phenylalanine strengthens its candidacy as a biomarker for future cardiovascular complications in MASLD.